The invention discloses a hazardous chemical transport vehicle illegal driving behavior detection method and system based on visual collaboration. The method comprises the steps that firstly, a data set is made, and image enhancement and normalization processing are carried out; marking dangerous driving behavior sensitive articles, and training the data set for the sensitive articles through a target recognition neural network; reading each frame of the video, detecting a sensitive object through the trained network model, and then performing human body posture key point detection on the frames; coordinate positions of the left and right hand key points in the frame image are obtained through human body posture detection, position information of the key points and the sensitive object recognition frame is judged, and corresponding illegal driving behaviors are recognized; and early warning is performed when the continuous time window of the illegal driving behavior exceeds a set threshold, and the video clip is reserved. The method can adapt to real-time detection and early warning of dangerous driving behaviors of a driver in a monitoring video, facilitates the timely discovery and prompt of non-standard driving behaviors, reduces the occurrence of traffic accidents, is various in detection types, and is good in recognition effect.