Abstract: Limited by the measured meteorological parameters, the accuracy of the traditional tropospheric model using standard atmospheric parameters is not high. The precise tropospheric model using parameter estimation method increases the number of parameters to be estimated of the observation equation and affects the convergence rate. For the absence of measured meteorological parameters, a fusion tropospheric model is proposed. Two non-measured meteorological parameter models are used to calculate the meteorological parameters at the average sea level and at the station respectively, and then the Saastamoinen model empirical formula is used to solve the zenith tropospheric delay. Precise point positioning(PPP) experiments were performed on RTKLIB software. The proposed fusion tropospheric model is free from the limits of measured meteorological parameters. The results show that when this fusion troposphere model is used, the positioning accuracy in the east, north and up directions is improved by 16 mm、1 mm、2.2 mm, compared with the Saastamoinen model, respectively.And, the positioning accuracy in the east, north and up directions is improved by 13.8 mm、0.7 mm、1.6 mm compared with the MOPS model,respectively. The positioning accuracy in the east, north and up directions is improved by 2.9 mm、0.4 mm、0.7 mm compared with the GPT/UNB3m+Sa model,respectively. The positioning accuracy in the up and north direction is close to the parameter estimation model. The PPP positioning accuracy is improved by using the proposed fusion tropospheric model.