In view of the problems in the existing research on enterprise performance evaluation, such as the single dimension of the evaluation index system, the short timeliness of the evaluation results, and the lack of discussion on the future performance level of enterprises, China's listed logistics enterprises was taken as an example to construct the performance evaluation index system of China's listed logistics enterprises from the financial and non-financial perspectives. The weight of each evaluation index and the expected performance value of each sample enterprise were determined by entropy-VIKOR algorithm. At the same time, an enterprise performance evaluation and prediction model based on AGA-BP neural network was constructed through the optimization of BP neural network using adaptive genetic algorithm (AGA). Finally, based on the data of 36 sample enterprises, the model was trained and tested. The test results show that the performance prediction model of listed logistics enterprises based on AGA-BP neural network is effective and practical.