Regarding the issue of insufficient attention to artificial intelligence methods in current evaluation of water resources intensive and safe utilization, an evaluation index system for water resources intensive and safe utilization was established. The nutcracker optimization algorithm(NOA) was employed to optimize the parameters of the support vector machine (SVM) model, leading to the construction of an NOA SVM model for the evaluation of intensive and safe utilization of water resources. By altering the proportions of training set and testing set, the evaluation result of the NOA SVM model was compared with that of the standard SVM model, and the performance of the NOA SVM model under different kernel functions was analyzed and compared. Furthermore, an empirical analysis was conducted on the 31 provinces (autonomous regions and municipalities) of China excluding Hongkong, Macao and Taiwan, to verify the effectiveness of the model. The results indicate that the NOA SVM model achieves a higher evaluation accuracy than the conventional SVM model in the evaluation of the intensive and safe utilization of water resources. In particular, the radial basis function kernel maintains a stable goodness of fit under different proportions of the training set, and its corresponding error metrics including MSE, RMSE, and MAE are all desirable and superior to those of the linear kernel and polynomial kernel. The top seven provinces (autonomous regions and municipalities) in the evaluation results of intensive and safe utilization of water resources are Tianjin, Hebei, Beijing, Henan, Shandong, Shanxi, and Inner Mongolia, and their core advantages lie in controlled development intensity, improved water use efficiency, and well adapted industrial structure. The bottom seven provinces (autonomous regions, municipalities) are Jiangxi, Jiangsu, Hubei, Shanghai, Hainan, Guangxi, and Xizang, and their core issues include excessive resource development, low water use efficiency, and insufficient ecological water use.