Abstarct: In this thesis, an algorithm for tracking a flying object is proposed baxsed on using joint color_texture feature to represent a target. The Mean Shift algorithm is used for object tracking. In addition to the typical color feature, texture feature such as average feature and local binary pattern are taken into consideration for object identification. Moreover, a Support Vector Machine (SVM) is being used in order to distinguish between an object and its background. The task of learning SVM is accomplished using forty photos that objects and backgrounds separated manually, each of which from different angles. It is imaginable that some possible backgrounds were not included in training data during the tracking phase. In this case, to alleviate the problem in hand, background extension has been applied. As long as there are a great number of samples for training the SVM, the data which does not play a role in determining the hyper-plane has to be eliminated, and the training phase should be commenced afterwards. After receiving the first frxame, one can test it with the SVM and express it in a frxamework subsequently, as well as using Mean Shift algorithm for tracking that frxame in the next ones. The simulation results confirm the superiority of the proposed algorithm in both accuracy and speed over the traditional methods.