Abstract The standard cuckoo search algorithm (SCSA) is an intelligent population optimization algorithm, which is also a heuristic search algorithm. The advantages of the SCSA (such as its convenient operation, heuristic searching, etc.) make it easy to find an optimal solution and maintain a wide searching range. However, the SCSA also has some drawbacks, such as long searching time, and the ease of falling on a local optimum. In order to solve the problems existing with SCSA, in this paper, an improved standard cuckoo search algorithm (ISCSA) was studied, which includes chaotic initialization and a Gaussian disturbance algorithm. As a case study, taking economic, social and ecological benefits as the objective function, multi-objective water resources optimal allocation models were constructed in Xianxiang Region, China. The ISCSA was applied to solve the water allocation models and a multi-objective optimal water supply scheme for Xinxiang region was obtained. Water resources optimal allocation schemes for the planning level year (2025) for 12 water supply sub-regions were predicted. A desirable eco-environment and other benefits were achieved using the studied methods. The results show that the ISCSA has obvious advantages in the solution of water resources optimal allocation and planning.