In the present study a GUI tool for the prediction of building energy performance based on a three-layered BP neural network and MATLAB was developed. The inputs for this tool are the 18 building envelope parameters. The outputs are building heating, cooling and total energy consumptions and the energy saving rate. Compared with the complicated mathematical equations, this tool provides a very easy and effective method for students to learn the effects of building envelope performance parameters on the building energy performance. Thus, this tool can be used in building physics and building energy efficiency courses for the design of energy efficient building.