Abstract: As considering the limitation both in huge amount memory space and the speed serial processing of the large-scale flow of data,this paper aims at dealing with the accuracy and data compression after the process.In parallel platform,by taking advantage of threshold sliding window technology,the flow of data is led into each processor in sections.In addition,function mining of the data model comes true under the help of function found algorithm of the Gene Expression Programming(GEP).As mentioned above,proposing the multiple data compression algorithm based on Gene expression programming for Parallel Function Replace algorithm PFR-GEP(Parallel Function Replace-GEP).The results on PC cluster shows that the algorithm effectively improve the compression ratio and computational speed,and has a linear speedup.