Nowadays, clinical decision support system become a part of daily life. Accurate diagnosis of liver cirrhosis helps in avoiding medical problems which may lead to death. The aim of the study is to build a fuzzy expert system for the diagnosis of liver fibrosis-stage (DLFS). The system uses machine learning tools and data mining statics to discover fuzzy rules, which help physicians to provide a fast and accurate diagnosis. The experimental have been performed on real dataset from clinical data sheets for 119 patients infected by chronic HCV. The evaluation results showed that the system identify liver fibrosis-stage with high degree of accuracy 95.7% and may decrease the need for liver biopsy.