Regional water safety systems are affected by social, economic, ecological, hydrological and other factors, and their effects are complicated and variable. Studying water safety systems is crucial to promoting the coordinated development of regional water safety systems and anthropogenic processes. Thus, a similarity cloud model is developed to simulate the evolution mechanisms of fuzzy and complex regional systems of water security and overcome the uncertainty that is associated with the indices that are used in water safety index systems. This cloud generator is used to reciprocally transform a qualitative cloud image with a quantitative cloud characteristic value, and the stochastic weight assignment method is used to determine the weight of the evaluation indices. The results of case studies show that Jiansanjiang's water safety systems were in a safe state in 2002–2011, but the water safety systems in the arid area of Yinchuan City were in a dangerous state in 2006–2007 because of climate factors and a lack of effective water and soil resource protection. The experimental results are consistent with the research subjects' actual situations, and the proposed model provides a tool for decision makers to better understand the security issues that are associated with regional water safety systems.