Moving object detection is one of the key tasks in image processing due to its important role in many applications nowadays. In spite of the fact that many relevant active researches exist, these methods still have drawbacks. This paper proposes an integrated method to detect moving object in video sequence frames. To make the detection more robust, we adopt two well-known methods namely, optical flow and active contour, and added a new adaptive spatial thresholding. The optical flow is first applied on two sequence frames to differentiate the shift of flow vectors of these two. Then, the lost areas are retrieved by our adaptive spatial thresholding. Depending upon the intensity and spatial correlation of individual pixel, they can improve the accuracy and robustness of segmentation process. In the final step, active contour converges to unite moving areas into one. The experimental results show that our method outperforms the previous motion-based methods, especially in term of processing time.