OpenMP(Open Multi-Processing)can achieve an ideal parallel performance on shared storage system. MPI+OpenMP hybrid parallel programming model is closer to the architecture of multi-core clusters. The hybrid model on the multi-core cluster was executed. The test results showed that the hybrid parallel programming model could reduce 95% of the run time of the serial program. And its parallel speed-up ratio was more than 20. This paper shows the hybrid parallel model is an improvement of the pure MPI parallel system with better speed-up ratio and scalability.