Based on the application of wavelet analysis to conventional logs, through extracting the residual logging gradient, the influence of lithology to logging curves can reduced for highlighting the fracture signs. As a results, the fracture resolution from conventional logging is improved. On the basis of core observation and description, combined with FMI information, different types of fractures are identified using geostatistics method. An identification method of fracture network is established by multi-logging parameters. The proposed identification method of fracture network has obtained good results in the application of identifying the types of fractures in carbonate reservoirs of F Burial Hill in Shengli Oil Field.