Abstract: To accurately predict the productivity of CBM wells, a multi-level fuzzy mathematics method was used to establish a comprehensive evaluation model including 3 secondary evaluation parameters(i.e. resource conditions, development conditions, and transformation processes) and 10 tertiary evaluation parameters. Combined with the data of reservoir parameters and stimulation technologies of 15 coalbed methane target blocks in China, the comprehensive evaluation work was carried out using the model. Finally, the actual production of each block was collected to verify the evaluation results, and the corresponding relationship between the evaluation results and the production was obtained. The results show that, the average daily production per well in most blocks does not exceed 600 m³/d, and the evaluation score is relatively concentrated below 0.64, which is mainly due to the poor development conditions and unreasonable reservoir stimulation. The production per well has a good correlation with the comprehensive evaluation results, and the relationship between them is power function. According to the evaluation results, when the comprehensive score exceeds 0.660 6, the average daily production per well is expected to exceed 1 000 m³. The research results shows that, on the one hand, it is necessary to strengthen reservoir evaluation and optimize favorable areas; on the other hand, optimizing suitable reservoir stimulation measures, to maximize the production capacity of coalbed methane development.