As the next generation power system

the smart grid has achieved the goals of high efficiency

safety

economy

andenvironmental friendliness.However

the deep integration of information physical systems poses a new security risk of false dataattacks.Therefore

a false data attack detection method based on improved particle swarm optimization(PSO)to optimize radialbasis function(RBF)is proposed.Firstly

the mean clustering method is used to optimize the hidden layer center values of theRBF neural network to improve network performance. Then

by designing weight coefficients and learning factors

theconvergence performance of the PSO algorithm is improved to optimize the field basis width and weight values of the RBFneural network

thereby improving the detection performance of the RBF neural network anomaly data detection model againstfalse data attacks. Finally

compared with current methods such as PSO-RBF and artificial bee colony optimization RBF

numerical examples have verified that the detection algorithm proposed in this paper can improve the detection rate by at least 0.83%.