The climate data used for dynamic energy simulation of buildings located in urban areas are usually collected in meteorological stations situated in rural areas, which do not accurately represent the urban microclimate (e.g. urban heat island effect), leading to potential simulation errors. This paper aims at quantitatively evaluating the effects of heat island on a high-rise building energy performance, based on the microclimate simulation tool ENVI-met and building energy simulation tool COMFIE. However the computation of microclimate is time consuming; it is not possible to simulate every day of a year in a reasonable time. This paper proposes a method that generates hourly "site-specific climate data" to avoid long microclimate simulation time. A coupling method of ENVI-met and COMFIE was developed for more precise simulation, considering the effects of heat island on buildings energy performance. It was applied to a high-rise building in Wuhan, China. The results showed that the yearly average urban heat island effect intensity at the height of 3 m was estimated to be 0.45 °C and decreased with the height. Compared to the simulation considering the outdoor temperature variation with the height and orientation, using the original climate data collected in rural areas led to an overestimation of the heating load by around 3.5 % and an underestimation of the cooling load by around 3.9 %. Compared to the weather file at the height of 3 m in the south neglecting the temperature variation along the height, the heating load was overestimated by 5.4 % and the cooling load was underestimated by 6.9 %. The methods proposed in this paper can be used for a more precise application of urban building energy simulation.