The algorithm proposed in this study estimates the multipath elevation of targets at low altitudes, including near-zero elevation. Although the double null algorithm is a maximum likelihood elevation estimation algorithm, it diverges when the target elevation is near zero. Addressing this issue, the selective double null algorithm improves the accuracy of multipath elevation at low altitudes by applying previously measured antenna near-field patterns. However, it cannot be applied to active electronically scanned array radars, since antenna beam patterns vary with the steered azimuth and elevation angles. In this paper, the proposed algorithm modifies the double null algorithm to estimate the elevation in the divergence state more accurately. Since it is based on the multipath energy function, the antenna near-field pattern is unnecessary. In addition, to increase the accuracy of elevation estimation, an operation method that selectively uses the estimated elevations according to several conditions is proposed. The proposed algorithm was verified through simulations.