The flamelet/progress variable (FPV) approach is further extended for a better prediction of NO emission in pulverized coal flames. The predicted results are compared with detailed chemistry simulation solutions for model evaluation. The results show that the original FPV approach with a conventional progress variable consisting of major species can predict the gas temperature and major products well but can roughly predict the NO variation. Therefore, the extended FPV approach with a modified progress variable including both NO and major species is further evaluated. The results indicate that the prediction accuracy of NO is sensitive to the proportion of NO introduced in the modified progress variable. The best prediction of NO can be obtained with the modified progress variable where the concentration of NO is comparable to that of all major species. However, the predictions of other quantities of interest are simultaneously deteriorated to some extent. Based on these analyses, a hybrid FPV approach is developed in which two flamelet libraries are constructed with conventional and modified progress variables, respectively. The results indicate that the hybrid FPV approach can predict both NO and other quantities of interest quite well at the same time under various operating conditions of strain rate, coal particle diameter, coal mass flow rate, and initial temperature.