A robust and low-complexity parameters estimation method, which combines TLS-ESPRIT and sparse signal representation (SSR) to estimate the parameter of coherently distributed sources is presented. The central direction-of-arrivals are estimated by TLS-ESPRIT based on approximate rotational relationship between two identical uniform linear arrays (ULAs), and angular spreads are given by basis selection method using SSR. Hence the method is a low-complex estimator that does not need a search in one or two-dimensional parameter space. And it owns more robustness in estimating angular spread than one-dimensional distributed source parameter estimator (DSPE). In addition, our method can estimate parameters of the sources with different angular signal distribution functions, which DSPE cannot work. The simulation results prove the effectiveness of our method.