In order to optimize the water resources allocation in Taiyuan City, water supply and demand scenarios considering water-saving awareness and unconventional water resources were created, the NSGA-Ⅱ+ARSBX algorithm was used to solve the optimal allocation model of water resources, and an evaluation model of multi-system coupling coordination degree was constructed to evaluate the solving ability of this algorithm. Based on the results of water supply and demand prediction, appropriate water-saving awareness scenarios were selected according to the available water supply under different guarantee rates. In the allocation results, the average proportion of water supplied by unconventional water accounts for 26.08%, which is only lower than the proportion of water supplied by external transferred water (35.77%). The coordination degree of the NSGA-Ⅱ+ARSBX algorithm in different level years under different guarantee rates are more than 0.8, which is better than that of the NSGA-Ⅱ algorithm. Studies show that medium or even high water-saving awareness is needed to improve water resources shortage under the guarantee rate of 75% and 95%. Furthermore, unconventional water resources account for more than a quarter of the total water supply, and incorporating unconventional water resources into the water supply system can play a significant role in alleviating water resources scarcity. Compared with the NSGA-Ⅱ algorithm, the NSGA-Ⅱ+ARSBX algorithm has better feasibility in optimal allocation of water resources and can provide a new approach for solving the optimal allocation model of water resources.