The skin-friction coefficient which indicates the degree of the stratum damage and the loss of production is important for evaluating reservoir contamination. A skin-friction coefficient prediction model based on PSO-BP neural network is presented in this paper, which integrates PSO and BP algorithm and takes full use of the global optimization of PSO and local accurate searching of BP. The examples of skin-friction coefficient prediction show that the prediction model works with quicker convergence rate and higher forecast precision, and can be applied to evaluate the degree of reservoir contamination effectively.