ue to ionospheric delays caused by the receiver code biases in the traditional multi-GNSS precise point positioning (PPP) using raw observations, the estimates can be negative values. An improved model of Global Positioning System/BeiDou Navigoction Satellite System (GPS/BDS) PPP with receiver differential code bias (DCB) parameters for raw observations is proposed in which receiver code biases on the first frequency of each system are constrained to zero and receiver DCB parameters are estimated. Ionospheric delays and receiver code biases are separated by the presented model. Additionally, the singularity between receiver clock offsets and ionospheric delays is reduced. GPS/BDS data from 4 stations of the Multi-GNSS experiment (MGEX) network are processed in static and kinematic modes. The results show that with the proposed PPP model, the average positioning accuracy and convergence time in static/kinematic mode are improved by 29.3% and 15.7%, 29.8% and 21.6%, respectively, in comparison with the traditional PPP model.