This paper describes a new model for predicting natural interior illumination in the urban context. A simplified radiosity algorithm is used to define the external luminous environment and this model structure is used to predict internal illumination from sky and external obstructions for the range of sky conditions. This illumination is predicted with comparable accuracy to the ray tracing program RADIANCE, but at a computational cost several orders of magnitude lower. A standard diffusing sphere approximation (the BRS split flux equation) is currently used to solve internally reflected light. Comparisons suggest that the accuracy of this approach is adequate for the purpose of predicting the energy implications of photoresponsive lighting control. The model is in a form readily amenable for inclusion into other programs that require internal illuminance as an input, to improve the accuracy of their predictions without recourse to external simulation programs.