Aiming at the problems of long optimization time, large empty distance and low optimization performance in the current zoom laser cutting path optimization method, this paper studies the zoom laser cutting path optimization based on machine vision. In this method, the parameters of laser cutting path are obtained according to the principle of machine vision. The relevant paths are traversed through computer technology. According to the path constraints and objective optimization function, the path optimization model is established, and then the model is solved based on the improved ant colony algorithm to obtain the best zoom laser cutting path and realize the path optimization. Experimental results show that when using this method to optimize the zoom laser cutting path, the optimization accuracy is higher than 95%, and the optimization time is kept at about 3.5s with the increase of test times. Therefore, this method has low optimization time, small idle travel and high optimization performance.