A dog pulls because pulling has always gotten them where they wanted to go faster. Stop moving the instant the leash goes tight, wait for the dog to return to your side or the leash to loosen, then continue walking, and the dog gradually learns that a loose leash is what actually gets them forward. A fixed-length leash paired with a well-fitted harness or martingale collar makes this far easier to teach than a retractable leash, which hands out more distance every time a dog pulls against it.
Stop the Pull, Not Just the Dog
Teach a loose leash by making sure pulling never gets your dog where they want to go.
petstore.comThe stop-and-wait method, step by step
Equipment that fights the training vs equipment that helps it
Retractable leash: extends as your dog pulls, handing out more distance as a reward
Fixed 4-6 foot leash with a martingale collar or well-fitted harness: stays consistent, easy to read
Hold the leash with a thumb lock
Loop the leash handle over your thumb instead of around your wrist. It gives you a faster, more secure grip if your dog changes direction suddenly.
Every walk your dog has ever taken has quietly trained them to pull
If pulling has ever gotten your dog to the tree, the other dog, or the door faster, they have already learned that pulling works. Dogs repeat what pays off, and a tight leash that still drags you forward is a payoff every single time. The fix is not a stronger arm. It is making sure pulling stops paying.
That is the entire logic behind the stop-and-wait method. The moment the leash goes tight, you stop walking completely and just stand there. When your dog looks back, slows down, or the leash slackens on its own, you continue. Repeated over and over on the same walk, a dog starts to notice that forward motion only happens on a loose leash.
Timing the reward matters more than the treat itself
The instant the leash goes slack, mark that moment with a word or a click and follow with a reward, whether that is a treat, a bit of praise, or just continuing the walk. Reward too late and the dog connects the treat to whatever they happened to be doing a few seconds after the good behavior, not the behavior itself. Early sessions are short and boring on purpose, done in the yard or a quiet street before adding the distractions of a real walk.
Equipment changes the physics of the problem too, in a way that has nothing to do with training skill. A retractable leash extends as your dog pulls against it, which means the leash itself hands out more distance as a reward for pulling, working directly against whatever you are trying to teach. A leash that stays a fixed length, paired with a properly fitted collar or harness, removes that built-in reward and makes the tight-leash moment easier for a dog to actually notice.
Consistency across every walk beats a perfect training session once a week
A dog who gets to pull on the evening walk after a well-behaved morning walk just relearns that pulling still works sometimes, which is enough to keep the habit alive. Every walk during the training period needs the same rule: tight leash means the walk pauses. Once loose-leash walking is solid in low-distraction settings, gradually add busier sidewalks, other dogs, and off-leash-adjacent areas so the skill holds up outside a quiet backyard.
The stop-and-wait method is the ASPCA's actual recommended technique, not a shortcut version of it
The ASPCA's virtual pet behaviorist guidance on leash pulling describes the same core method used here: when a dog pulls, stop walking and wait until the dog returns to your side or the leash loosens, then continue on. It also recommends holding the leash with a thumb hooked through the handle loop rather than looped around the wrist, which gives a handler a faster, more secure grip if a dog lunges or changes direction suddenly.
On equipment, that same guidance points to a martingale collar as a solid option for dogs that do not pull hard and have no medical reason to avoid neck pressure, with a well-fitted harness as an alternative depending on the individual dog. This article's point about retractable leashes working against training is a mechanical fact about how those leashes function, not a claim pulled from the approved source list, and is noted here for that reason.