Abstract: A new trust evaluation algorithm based on particle swarm evolution in cloud transaction is proposed. Under the O2O cloud model, the network structure and trust attribute data analysis model is constructed, and the influence parameters are constructed by using the method of particle swarm optimization. The feature value is extracted from the trust property and the control system is realized. The results show that this method has high accuracy and good real-time performance, and has good real-time performance, and has a good ability of online filtering and analysis.