The purpose of this study is to use T-GM(1,n) and T-GM(2,n) to predict the number of teachers for admission in Taiwan. T-GM(1,n) and T-GM(2,n) are two prediction models based on Taylor approximation method in grey prediction to improve the predicted accuracy of GM(1,n) and GM(2,n). Two combined models above can obtain the most optimal values of prediction by multi-times approximate calculation. In addition, researchers have used the MATLAB software to build a MATLAB toolbox for two prediction models which is more convenience to calculate and process the predicted outcomes. The results in this study show that, they are not only conducted to serve as a reference for the educational administrators but also can assist the government in developing future policies regarding educational management. The predicted data are required high accuracy and reliable to contribute to the success in the educational development of the country.