Abstract: In some techniques, such as smart antenna and wireless location etc., it is very important to track accurately the direction-of-arrival (DOA) of aim signal. Most conventional DOA tracking algorithms are based on the assumption that the aim signal is point sources. However, in dense mobile communication environments, signal multipath scattering will result in angular spreading of the source energy, a distributed source model may be more appropriate. Hence, it is more significant and valuable to research the DOA tracking technique of distributed sources. Aiming at coherently distributed sources, this paper presents a fast algorithm for the central DOA tracking by using uniform linear arrays (ULA). The presented algorithm firstly updates the signal subspace in real-time by adopting the orthonormal projection approximation subspace tracking with deflation (OPASTD) algorithm, and then a low-complexity DOA method is used to estimate the central DOA of distributed sources. The algorithm avoids the decompositions of the sample covariance matrix which must be repeated in the classical DOA estimation algorithms and has lower computational complexity, and it provides excellent tracking performance and can track effectively multiple sources. In addition, without the prior knowledge about the angular distribution shapes of distributed sources, the robustness of our algorithm is good. Simulation results validated the efficiency of the presented algorithm.