A cross sectional analysis for residential water demand was conducted to help understand and explain the spatial and temporal variations in per capita water use in the rapidly growing city of Riyadh, Saudi Arabia. The analysis was based on data previously collected from May 1983 to June 1984. 195 randomly selected households were distributed to three groups according to house condition, household income level, and social and cultural factors. The generated models using stepwise multiple regression indicated that plot size and number of males, females and children are the most significant independent variables. Although, coefficients of determination achieved for most of the developed models were low (0.2-0.5), the independent variables could still explain a part of the variations fur such a complex social and cultural structure.