Predicting the DNA splice sites has become the most attractive and important issue in the field of genome information research due to its great help to find gene coding area.Current gene recognition algorithms mainly consider the global features of the coding area,instead of the specific information of the splicing sites,which are usually unable to recognize the splicing sites accurately.A new method based on the conditional random fields (CRFs) is proposed for splice sites prediction.CRFs can capture long distance dependent relationships between labels and avoid label bias.Experimental results show that CRFs is suitable for splice sites prediction and can improve performance.