Aiming at the existing issues of the shortage, unreasonable utilization and incomplete allocation theories of water resources, an intelligent method of regional water resources allocation is investigated on the basis of data-driven approach in order to improve the rationality of utilizing regional water resources and provide technical support for the planning and management of the regional water resources. Firstly, a framework, made up of the decision-making model, scheme evaluation model and plan base of the water resources allocation, is presented. The evaluation model is structured with the evaluation index system of the water resources allocation based on analytic hierarchy process and the fuzzy evaluation method. On one hand, it preliminarily examines the allocation scheme from the decision-making model or after adjustment; on the other hand, it provides allocation plans to the base as the decision-making model’s training samples. The decision-making model is built on the basis of artificial neural network being trained according to the plan base, which is used to provide an initial allocation scheme. The effectiveness of the framework is verified using the relevant data of 2010 from Jiaxing city. The results indicate that the intelligent allocation framework of the regional water resources based on data-driven approach can generate the appropriate allocation schemes of the water resources based on two strategies of using water in Jiaxing city, i.e. harmoniously developing and considering both regional and overall water development with a constant amount of water resources. And as an effective support tool for the planning and management of the regional water resources, the scheme is of self-adaptation, fault-tolerance and intelligence.