In this research, a preliminary mixed multi-frequency PPP solution strategy is analyzed and tested based on the combination of BDS-3 and GNSS observations. Firstly, the multi-frequency observations are combined and its coefficients are rapidly estimated by least square; then, the inter-system bias parameter and the stochastic model are introduced into the function model; finally, the mixed PPP solution and its software are developed and verified by three groups of experiments. According to experimental results of 96 stations and ten-day MGEX observations, it is indicated that the root-mean-square error (RMS) of positioning and the convergence time are significantly optimized with the aid of additional frequencies, where the accuracy improvements of multi-frequency and multi-GNSS scheme in east (E), north (N) and up (U) directions can respectively reach up to 23.2%, 13.3% and 23.8% compared with traditional BDS-3 dual-frequency ionosphere-free (IF) PPP model; and the corresponding convergence time is reduced from 18.54min to 13.18min. Meanwhile, from the results of multi-frequency BDS-3 PPP experiments, it is suggested that a better performance of positioning and convergence can be obtained, where the position RMS of E, N and U directions are reduced with 38.2%, 23.9% and 26.3%, and the convergence time is decreased from 23.86min to 12.43min for BDS-3 combined all of observations, compared with BDS-3-only solution. Furthermore, in the vehicle experiment of multi-frequency kinematics PPP, a convergence process can be found for different scenarios. Moreover, residuals series are different from each solution, in which the reducing with 71.1%, 33.3% and 77.1% in directions of E, N and U can be obtained compared with traditional BDS-3 dual-frequency IF model in kinematics experiments based on multi-GNSS and multi-frequency scenario. Therefore, it is meaningful to recommend the mixed PPP solution in GNSS community to fully use the multi-frequency and multi-GNSS observations by the adaptive combination of different observations.