This paper summarizes the results of a comparative analysis for four models utilized to predict solar radiation in Kuwait City, Kuwait. The four models include the Kasten model, Zhang and Huang model, Muneer model, and neural network based model. The analysis was based on hourly measured solar data for Kuwait City. The measured hourly solar radiation data are obtained for the year 1994 and include global, direct, and diffuse solar radiations. Nonsolar weather data for the same year and site are obtained from the US National Climatic Data Center (NCDC). Weather files suitable for building energy simulation are developed using measuered data as well as predictions from the four solar models. A series of simulation analysis to determine the impact of solar model selection for the weather file on the energy uses predictions from a whole-building simulation program using office buildings in Kuwait. The results of the validation analysis and the simulation evaluation indicate that Zhang and Huang model is suitable for the predicting hourly solar radiation suitable for energy analysis of buildings in Kuwait.