To solve the limitations of traditional regional communication network node identification, we used DS evidence combination method to integrate different node attributes to overview the criticality of regional communication network nodes. First we divided influencing factors of key node evaluation into different attributes such as node command level, node information flow and node topology characteristics, so that the key node evaluation system of the regional communication network can be established, then we calculated node attribute weights by using analytic hierarchy process. Verification by simulation and compared with traditional methods, the results show that the method can overcome the shortcomings of single attribute evaluation, and it has more detailed evaluation ability and get more accurate evaluation results.