Abstract: The emergence of big spatio-temporal data brings brand new perspectives as well as challenges for us to investigate and understand urban space. Due to existence of GPS position error, it is inevitable to adopt the map-matching methods to map the spatio-temporal trajectories onto geographic space. This research focuses on the low-sampling trajectories of floating cars in urban road networks by formalizing the map-matching process and exploring the influence of both the geometric and topology constraints on matching results. To solve the problem of matching low-sampling GPS data in the context of complex urban road networks, this paper proposes a topology-constrained incremental matching algorithm (TIM). Utilizing a sample GPS trajectory of Beijing float car as an example, the TIM algorithm is verified to be efficient and accurate give various road network complexity. Our study is valuable for the pre-processing of massive spatio-temporal data, and has the potential to benefit trajectory data mining and related urban informatics research in the future.