Traditional precise point positioning (PPP) based on undifferenced ionosphere-free linear combination of observations has many advantages such as high accuracy and easy operation. PPP usually uses the Kalman Filter (KF) to estimate state vector. However, the positioning performance depends on the accuracy of the kinematic model and initial value. The inaccurate kinematic model or initial value will lead to filter performance degradation or even divergence. To overcome this problem, this paper proposes a PPP method with an additional baseline vector constraint, which uses the direction information and length information of the baseline to correct the estimated position of the receiver. By reducing the error covariance matrix of the float solution, the algorithm improves the accuracy of the float solution. By using the collected real GPS static and kinematic data, the performance of the traditional model and the proposed model in this paper are compared. It is shown that the additional baseline vector constraint improves the PPP solution effectively in comparison with that of traditional PPP model. Additionally, the contribution of the additional constraint is up to the accuracy of the prior information.