Lie recognition is a significant part of Criminal Investigation, and eye movement technologycan be used as a powerful means to identify lies because of its high-precision and non-invasivecharacteristics. By constructing a simulated theft crime scene eye movement experiment and comparingthe eye movement data of different subjects on target and non target stimuli, it was found that:(1) when “guilty” and “innocent” subjects of the same age face the same stimulus type, there aresignificant differences in the eye movement data of three groups: fixation time, eye skip distance, andpupil diameter, resulting in a separation effect; (2) When facing different types of stimuli, “guilty” subjectsshowed significant differences in the first fixation time, total fixation time, eye skip distance, average pupilsize, average fixation time at all fixation points, re fixation ratio, and return time of seven groups of eyemovement data, and also showed separation effects. This result provides a new enlightenment for eyemovement lie detection: by integrating various eye movement indicators, and considering multiple factorssuch as subjects' attitudes, stimulation order and environment, it can effectively distinguish suspect andinnocent people to a certain extent, thus effectively promoting the development of case investigation.