As global water scarcity becomes increasingly acute, water demand forecasting has emerged as a critical component in water resource management and planning. This review aims to comprehensively survey and analyze the current state of research, existing issues, and development trends in the field of water demand forecasting. Presently, there are numerous studies on water demand forecasting; however, most of the forecasting results tend to be overestimated. On the mechanistic level, research has gradually shifted from considering single factors to accounting for the complex influences of multiple factors. This paper summarizes the mechanism of water demand from the three levels of agriculture, industry, and residential life. In terms of forecasting methods, various techniques have been explored and applied, particularly new methods based on artificial intelligence and machine learning, which have demonstrated significant advantages in improving forecasting accuracy and handling nonlinear relationships. Despite the notable progress and practical achievements in water demand forecasting, several challenges and issues remain. Future research should focus on diversifying methodologies, comprehensively considering multiple influencing factors, further refining forecasting models and technical systems, strengthening uncertainty and risk management, and emphasizing practical applications and policy guidance.