In small cities, reliable potable-water demand forecasting is essential for designing water supply systems. This study compares several demand-forecasting approaches to identify the one that best reproduces historical observations. The analysis focuses on a case study with consumption data from 2017 to 2024 and evaluates four methods: (1) forecasting based on population and per-capita supply; (2) forecasting derived from the average number of inhabitants per customer; (3) forecasting using monthly consumption data; and (4) forecasting using total annual consumption. All methods draw on historical customer-level consumption records, are applicable regardless of climatic conditions, and were fitted with multiple regression models, selecting the specification with the highest coefficient of determination (R²). Forecast performance was assessed with statistical indicators -mean absolute error and root mean squared error- revealing that the user category forecasting method with monthly data exhibited the smallest error margins relative to the historical series, making it the most accurate and consistent tool for estimating future potable-water demand in comparable urban settings.