It is only a tentative attempt to analyze motion trajectory data in the data mining style. This paper proposes an algorithm based on Apriori, a conventional data mining method, for mining frequent trajectory patterns of moving objects from surveillance video. Firstly, feature points are extracted for presenting continuous trajectories. Then a trajectory similarity measure is designed to obtain the frequency of trajectories. Finally, based on the values, Apriori algorithm is applied to automatic finding of interesting patterns in 2D object trajectories. Experiments on real life and man-made trajectory data set show the validity of the algorithm.