Predicting the temporal evolution of the demography and the residents’ spatial movements would immensely aid the estate development and urban planning. The evolution of population in three townships of Singapore is simulated at neighbourhood scale using a novel agent-based probabilistic approach with inputs from large-scale survey and statistical data. The demographic changes due to age-dependent rates of death and fertility are studied by considering the inter-ethnic marriages that has a varying probability depending on the ethnicities of the male and female partners. The predicted changes in the age and household compositions and family types have been found to reflect the population trends in Singapore over the past years. The decline in family types that contain children and the structure of age composition over years underline the issue of prevailing low fertility rates. The strategies for incorporating the population relocation to consider the long-term spatial movement are also discussed. In Singapore’s context, we consider in the relocation model an added complexity of ethnic quota for the residential units developed by public housing board. The ethnicity dependent parameter coupled with other parameters that represent the number of children in a household besides their size, the household income, the proximity of children’s schools, and the places of employment could play a strong role in predicting the spatial evolution of the residents. These predictions can be used by the urban planners and policy makers to improve the quality of life in Singapore.