Abstract: It’s economical and simple to use Fuzz Comprehensive Evaluation Method to recognize big channels, but the selected index and established evaluation system will directly affect the accuracy of results. According to the actual reservoir characteristics and the data availability, 9 static indexes that affect the form of big channels and 9 dynamic indexes that represent the form of big channels are selected as indexes of this fuzz model in this paper. And then selected indexes limits are determined based on the actual reservoir data. With the selected indexes and limits, fuzz pattern recognition model and evaluation system based on actual production situation are established. By comparing the results of fuzz recognition model and tracer interpretation, it figures out that the coincidence rate is up to 80%, and which mean the applicability and reliability of method established in this paper is very well.