Abstract: This paper proposed a three-level completion strategy for the passenger public transport travel chain that integrated bus and subway card swiping data to address the problem of missing data at the public transport stop. This completion strategy included three levels: transfer itinerary chain completion strategy, public transport travel chain completion prediction model, and maximum probability disembarkation prediction model. The paper utilized spatial position constraints in the same day and next day transfer scenarios to complete the conventional public transport drop off points, introduced an improved density based clustering algorithm DBSCAN（Density-Based Spatial Clustering of Applications with Noise), combined weighted geometric centers to construct an optimization model for public transport stop points. Based on the probability distribution of bus routes getting off, the paper constructed a maximum probability getting off model. Through this three-level completion strategy, the paper implemented the construction of a full process completion system for public transport stop points. The results of experiments based on real travel data show that the proposed completion strategy can implement a completion accuracy of 92.5% for the public transport travel chain, which is more than 10% higher than the accuracy of traditional algorithms. Moreover, over 80% of the error stations are concentrated within ± 1 station range of actual stations. It proves its superiority in accuracy and robustness.