Dissertation (MSc Statistics)Water is one of the fundamental element for life. For that reason, water usage forecasting is required for a proper planning and management. Because there is no single methodology that is used to forecast water demand (quantity of water to be used) across the world, this study intended to develop a statistical model for household water usage in Dodoma Municipality. Longitudinal and cross-sectional research design data from DUWASA, NBS and 395 households connected with piped water were used Model development involved two variables, population and quantity of water used. The study has found over an existing relationship between the population and the quantity of water used is non-linear. Quadratic model is the best model among non-linear models which fit well the data. In addition, findings indicate that, water price, income and household size are factors which significantly influence water usage in Dodoma Municipality. DUWASA is performing well by supplying water in line with the usage. Customers are satisfied with DUWASA customer care but dissatisfied with provision of information, water connection and supply likewise complains response. The overall DUWASA customer’s satisfaction stands at 54.4% which imply that, customers are moderately satisfied with DUWASA services. The study recommends that, water authorities may consider population, household size and head of households’ income, as variables in planning and management of water resources. To raise customer satisfaction, DUWASA needs to improve its services especially on factors which customers were observed unsatisfied.