The conjugate symmetry property of L-shaped ULA is applied to two received data matrices from two orthogonal subarrays of L-shaped ULA, and hence, make two new received-data-like matrices, which potentially increases the diversity of received data. In the following, four observations cross-variance matrices are formed from both the received data matrices and the received-data-like matrices. By developing the maximal overlapping of those observations cross-variance matrices, a new generalized received data matrix with the generalized aperture is constructed, which might not only improve the DOA estimation performance, but resolve the angle ambiguity problem. To this end, the efficient PM-ESPRIT-like method is used to estimate 2-D DOA without global search and angles pairing matching. Numerical simulations are offered to justify the performances of the proposed method.