Abstract: In order to describe the characteristics of RF power amplifier accurately, it is very important to establish a good behavior model in the simulation. In this paper, a kind of RF amplifier behavior of BP-RBF cascade neural network optimized by ant lion algorithm (Ant Lion Optimizer, ALO) is proposed. First, the RF power amplifier circuit is designed by Freescale semiconductor chip, and the voltage data extracted from the designed circuit is processed. Then, the weights and thresholds of the BP-RBF neural network are optimized by using the ability of several individuals in the population of the lion ant in parallel, and the improved optimized ALOBP-RBF neural network model is simulated by MATLAB. The accuracy of the model is verified by comparing the voltage root mean square error. The simulation results show that, compared with the BP-RBF and GABP-RBF models, the model has higher accuracy and faster convergence speed, and can accurately simulate the characteristics of the power amplifier, which is of great significance for the modeling of RF circuits.