This study introduces a multi-source data fusion model of the tropospheric delay over China and uses GNSS data, meteorological data, and Global Pressure and Temperature 2 wet (GPT2w) model data to derive the model coefficients. Fifty-one nationwide GNSS stations were selected to evaluate the accuracy of the fusion model through a detailed analysis of the model performance based on factors including the overall accuracy, seasonal accuracy, and impact of the station location. The results show that the multi-source data fusion model integrates the advantages of various individual models and thus has a higher accuracy and stability despite the occurrence of a certain degree of decline in accuracy in the individual models under certain conditions. A comparison with previous models developed using only GNSS data demonstrates that this fusion model improves the overall accuracy by 17.7%.