Abstract: To accurately estimate the direction of arrival (DOA) under conditions of small samples, low signal-to-noise ratio and high correlations of sources, according to the compressive sensing theory, DOA estimation based on weighted l₁ norm sparse signal representation was proposed by using the sparse distribution of the spatial signal source bearing. The l₁-SVD algorithm was improved by this algorithm, the receiving matrix was proposed, the weighting matrix was determined according to the orthogonality of the subspace, and the sparse vector was obtained by the optimization using the weighted 1 norm as target function for minimization. Then, the DOA of the signal was obtained. Simulations demonstrate that the proposed algorithm can effectively suppress the deviation by sparse signal reconstruction method under weighted processing l₁ norm. At the low SNR, DOA can be estimated accurately and steadily, and the precision of DOA estimation can be improved effectively.