Objective This study aims to scientifically evaluate the operational risk level of check dams on the Loess Plateau in Inner Mongolia, thereby providing theoretical support for risk warning and engineering management of check dams in this region.

Methods First, a multi-dimensional and hierarchical risk evaluation indicator system was constructed by identifying key influencing indicators using the decision-making trial and evaluation laboratory (DEMATEL) method. Second, the weights of the indicators were determined separately via the fuzzy analytic hierarchy process (FAHP) and the order relation analysis (G1) method, and game theory was applied for weight fusion to ensure the rationality of weight allocation. Finally, quantitative assessment was performed based on the cloud model. Specifically, the cloud characteristic values of the indicators were calculated using the backward cloud generator, and the comprehensive evaluation cloud was generated by combining indicator weights with the forward cloud generator. Additionally, cloud similarity was introduced to quantify the membership degree of the comprehensive evaluation cloud to different evaluation grades.

Results An empirical analysis was conducted with the Housha Check Dam in Inner Mongolia as the research object, and the results showed that the operational risk evaluation level of the check dam was “basically safe,” with a normalized membership degree of 70.5%. This result was highly consistent with those obtained using the fuzzy comprehensive evaluation method and the set pair analysis method and matched the actual engineering situation.

Conclusions This study confirms that the constructed evaluation system and method are effective and reasonable. It can accurately reflect the operational safety and risk status of check dams, and provide reliable references for risk management and control of check dams in similar regions.