Abstract: Risk evaluation of natural gas pipelines is ambiguous and uncertain; however, the existing methods couldn't realize the effective integration of qualitative concepts and quantitative descriptions, lacking uniform standards in the selection of risk indicators and the determination of weights, which urgently requires a systematic evaluation method. Based on the domestic and international natural gas pipeline accident statistics, 30 influencing factors in 6 aspects, including third-party damage, corrosion, environmental disasters, dis-operation, pipeline defects, and operational failures, were identified and constructed as evaluation indexes using a fishbone diagram method. Employing the game theory assignment method, the subjective and objective weights calculated by combining the G1 and improved CRITIC methods were comprehensively assigned. Matlab was used to determine the characteristic parameters of the cloud model and generate the cloud diagrams, and the affiliation degree and the comprehensive affiliation degree of each index under different risk levels were determined by the forward cloud generator and the level judgment was conducted. The four pipeline sections of the evaluated branch line were respectively medium, low, low, and lower risk, which coincide with the actual situation in the field and provide theoretical guidance for natural gas pipeline risk assessment. Results show that the proposed model effectively reduces the subjective uncertainty of the decision maker and eliminates the objective error between the data, performing highly practical and reliable.