When traditional technology is applied to the safety evaluation, it will be influenced by external interference factors, so that the evaluation is not accurate. It is necessary to introduce network entropy counter quantization technology to evaluate the safety of active defense network driven by big data, so as to improve the accuracy of evaluation results. Construction of Petri network security evaluation based on the mathematical model of the active defense principle, attack behavior reflects the model caused by indirect impact on the network, to reduce interference caused by indirect effects should be introduced to weigh the benefits in order to improve the network entropy, network security, by weighing the results of network active defense safety assessment. The experimental results show that the accuracy of the evaluation is up to 96%, and the user can use the network to ensure the security of the user"s personal information under the large data drive.