Abstract: With the advancement and integration of mobile computing technology and intelligent transportation systems (ITS), automatic traffic flow analysis has become an important research issue. However, current traffic flow statistical analysis methods, such as stationary sensor/camera based methods (monitoring from traffic sensors or optical devices), air/space borne methods (monitoring from airplanes or satellites), and floating-car based methods (monitoring from floating/probe cars), have a lot of limitations in terms of data sampling costs, data processing efficiency, and information analysis accuracy. To solve these problems, a new traffic flow analysis framework, network-constrained moving objects database based traffic flow statistical analysis (NMOD-TFSA) model is proposed, in this paper. Through an online statistical analysis mechanism based on the spatial-temporal trajectory data submitted by moving objects, NMOD-TFSA can get the real-time traffic parameters of the transportation network. By taking the topology of the traffic network into consideration, NMOD-TFSA can reduce the communication and computation costs in data sampling. Besides, in data analysis, NMOD-TFSA can provide better accuracy since the analysis is based on the precise spatial-temporal trajectories of moving objects. The experimental results show that compared with the floating-vehicle method, which is widely used in current traffic flow analysis systems, NMOD-TFSA provides improved performances in terms of communication and computation costs, flexibility, and accuracy.