How to Stop a Dog From Jumping on People (What Actually Works)

The jumping problem is not a mystery. You know why your dog jumps: it worked. Someone pet them, said "hi baby," pushed them down in a way that was basically a full-body interaction, or just made eye contact. All of that is a reward. The dog filed it away. And now here we are.

Why Dogs Jump (and Why Telling Them "Off" Doesn't Help)

Dogs jump to greet. It's social behavior, and in the dog world, getting close to faces is how greetings work. Your dog isn't being dominant or disrespectful. They're being a dog. The problem is that jumping on a 90-pound human is annoying at best and dangerous at worst, and most dogs have been accidentally trained to do it by the very people trying to stop it.

Every time someone says "off" and then gives the dog any attention, they've reinforced the sequence: jump, hear a word, get attention. The word becomes part of the routine. The dog doesn't process "off" as a stop command. They process it as a cue that attention is incoming.

Telling a jumping dog "no," "off," "down," or "stop it" while maintaining any eye contact at all is not training. It's narrating. The dog hears sounds, sees your face, gets attention. The behavior continues.

The Method That Works

Remove the reward entirely. The moment your dog's front paws leave the ground, you turn your back. Not after a second. Not after you say "off." Immediately. You cross your arms, look at the ceiling or wall, and say absolutely nothing. No eye contact. No pushing the dog away. No verbal correction. You become the most boring object in the room.

The instant four paws return to the floor, you turn back around, crouch down, and give calm attention. Petting, soft praise, whatever your dog responds to. You are marking and rewarding the four-paws-on-the-floor behavior, not the sit (yet), just the floor contact.

If the dog jumps again the moment you turn around, you turn away again. Immediately. This is the whole game: jumping predicts you become unresponsive, four paws predicts attention. Dogs are good at learning patterns. This pattern is learnable.

What to Teach Instead of Jumping

The ignore method stops the jumping, but it helps to give the dog an explicit alternative. A sitting dog cannot simultaneously jump. If "sit" is on a solid stimulus-response cycle, you can start asking for a sit at the moment of greeting and rewarding that instead.

The sequence looks like this: dog approaches, you ask for sit before the jump attempt, dog sits, dog gets greeting. Over time, the dog learns that approaching a person means sit, not jump, because sitting is what actually gets the interaction started.

This only works if the sit is already well-established in non-exciting environments. Practicing "sit" in the kitchen when nothing is happening is fine, but that's a different skill than sitting when a person walks through the door and the dog has been alone for eight hours. Train the sit in gradually increasing excitement levels, not just in calm conditions.

The Mistakes That Reset Your Progress

Inconsistency is the only reason this method fails. Not the dog's personality, not the breed, not stubbornness. Inconsistency. Here is what inconsistency looks like in practice:

You've been doing the turn-away method for two weeks and it's working. Then you come home tired and the dog is excited and it's kind of cute and you just let them jump a little. That single event tells your dog that jumping works sometimes. In behavioral terms, this creates a variable reinforcement schedule, which is the most persistent schedule there is. Slot machines work on the same principle. Your dog will now try harder and longer before giving up, because sometimes it pays off.

Other common resets: letting your dog jump on "certain" people (joggers, friends who say they don't mind), letting them jump with front paws only ("as long as they don't fully jump"), letting jump-adjacent behaviors go unmarked (paws on the thigh, standing up partway). The dog does not understand partial rules. If jumping sometimes gets attention, jumping continues.

The Guest Problem

This is where most dog owners give up, and it's the hardest part. You cannot control what strangers do on the street. You can brief guests before they come in. You can put the dog on leash when guests arrive so you control the greeting sequence. You can have guests turn their backs rather than saying hello to a jumping dog.

What you cannot do is train the dog out of jumping at home and then have guests squeal and hug the dog while it's in mid-air and expect the training to hold. The dog doesn't have a "jumping at home" category and a "jumping at guests" category. It just has "jumping sometimes gets rewarded."

When guests arrive, keep the dog on leash or behind a baby gate until the excitement settles. Then release and reward calm behavior. It takes about four minutes of management and then everyone can relax. This is easier than apologizing for a dog who just knocked someone's grandmother over.

How Long Does It Actually Take

With genuine consistency from every person the dog interacts with, most dogs show real improvement in two to four weeks. The jumping doesn't disappear on day three. There's usually an extinction burst first: the behavior gets worse before it gets better, because the dog tries harder when something that used to work stops working. This is normal. It means the training is working. Keep going.

Without consistency, the timeline is indefinitely extended. Six months of intermittent training produces a dog who jumps intermittently, because that mirrors what they've been rewarded for. The math is boring but accurate.

If you've been at this for months with genuine effort and nothing is changing, it's worth a session with a force-free trainer who can watch the specific dynamic, because sometimes the problem is in an interaction detail that's hard to catch on your own. The separation anxiety post has notes on when professional help makes more sense than continuing to DIY, and the principle applies here too.

Jumping is one of the most solvable dog behavior problems there is. The only thing that has to change is human consistency. Which is, historically, the harder training challenge.