Cities face pressing challenges from rapidly growing populations, a warming global climate, air pollution, resource depletion and energy scarcity. Computational models can assist in understanding interactions and planning solutions. This thesis documents a series of developments intended to improve the accuracy and capability of environmental models that represent urban land surfaces in weather and climate simulations. Developments focus on efficiently representing physical and behavioural processes at the neighbourhood to city scale. Major contributions and key findings include: 1) Development of a new heat conduction scheme which is more accurate than a method commonly used. Through a change in the discrete formulae representing conduction, the new scheme is better able to match exact solutions to heat transfer through typical urban materials, and reduces land-atmosphere flux errors when integrated within an existing urban land surface model. Improvements are achieved without increasing computational expense. 2) Development of a new building energy model that predicts neighbourhood-scale energy consumption based on weather conditions and building structure. The scheme includes important internal-external heat transfer processes, as well as a novel representation of human behaviours derived from a statistical model of the national electricity network. The scheme is integrated within an existing urban canopy model to capture dynamic feedbacks between energy use and the greater urban environment, and is able to reproduce observed diurnal and seasonal variability in building energy use. More complex physics-based processes are found to be beneficial only when human behaviours are appropriately represented. 3) Development of a new land-atmosphere coupled framework which simulates interactions between energy use, waste heat, urban and global climate. The new framework is evaluated and used to run 100-year simulations under the climate change projection scenario RCP8.5 to investigate how energy demand will change in a warming climate. With rising global temperature and increasing air conditioner use, peak electricity demand will likely surpass gas demand in Melbourne this century. These developments are integrated within a free, open source model: the Urban CLimate and Energy Model (UCLEM). Integrated modelling tools such as these will be instrumental in planning for better environmental, health and economic outcomes in cities.