As the breakage of old cement concrete in urban, the use of asphalt concrete overlay was more and more popular. Seldom actual test measurement datum and prediction and evaluation methods can be found. In order to scientifically and accurately predict the future composite pavement situation, evaluation indexes and main influence factors of composite pavement were analyzed. Then functional performance, structure performance, safety performance, and exterior qualities were selected as the evolution index. The two prediction models of BP network model and hybrid algorithm based on neural network and genetic algorithm were built respectively. Forecasting result shows that neural network model based on genetic algorithms coincides with BP neural network.