As one of the main error sources in the Global Navigation Satellite System (GNSS), the accuracy of tropospheric delay correction model influences the estimated coordinates, especially in vertical direction; therefore, it is necessary to make assessment of different tropospheric delay correction models. An assessment of four commonly used models, including EGNOS/UNB3m/GPT/GPT2, is made in this paper. The result shows that the RMS of the four models remains 4-5 cm, and the difference between each model is less than 1 cm, with GPT2 being the most accurate model; and the RMS of each model in summer is relatively much bigger than in winter, due to the abundant water vapor in summer, which makes it harder to model tropospheric delay precisely. All models show poor accuracy in southeast China, thanks to the rather changeable weather and the abundant water vapor there; and no model is sensitive to the height, since the accuracy varies little with respect to the variation of altitude. These findings provide a reference for the GNSS users in China area when choosing a suitable tropospheric delay correction model.