The propagation of faults in software can cause some functions to work incorrectly, and eventually leads to software failure. Identifying the critical nodes that have a massive impact on fault propagation in software, can help improve the design and optimize the structure of these nodes. At present, using complex network theories to analyze the software structure, and then identifying the critical nodes in software have become a new research hotspot. This paper proposes a method based on software dynamic execution network to identify the critical nodes of fault propagation. Firstly, we construct the software dynamic execution network and define the measure called “influence scope of propagation” to represent the critical level of nodes in software. Then, we select and combine some structure features possessing a strong correlation with the measure of “influence scope of propagatio”, and use the decision tree algorithm to construct the identification model for identifying the critical nodes. Finally, the validity of the identification method is verified by two instances.