The design and testing of a real-time system for tracking air-targets detected by TV camera is presented. The system is implemented using hardwired circuitry involving both analog and digital concepts. This study is a development of a previously implemented system, whereby adaptive thresholding was utilized to segment the image scene by separating the objects from the background. A target was enclosed in a tracking window to reduce the amount of data to be processed. The trajectory of the target was determined by enabling the window to track the centroid of the target image in successive TV frames. The tracking window is modified to have a size adaptive to the area of the target image to account for movement along the camera line of sight. Prediction capability is introduced by employing a quadratic predictive filter for estimating the location of the target image in the next frame. The filter facilitates the tracking of maneuvering targets, and provides a simple technique to track targets under occlusion. The performance of the system is evaluated by conducting several tests on computer and camera generated scenes for different target and background conditions. The tests reveal encouraging results in terms of accuracy and reliability, and show the effectiveness of the system in tracking targets in images containing high levels of noise.