Aimed at the problems that being affected by multipleaperture influence such as small array aperture, low angle measurement accuracy and poor performance, etc., large angle measurement error still remains in spite of the fact that the existing low elevation estimation algorithms mostly are mostly used by the uniform linear array (ULA) signal model for echo reception and processing, and then use the superresolution algorithm for low elevation estimation, are approximately feasible, and the performance is not good, a real valued low elevation estimation method based on physical array of CPA is proposed. Firstly, the signal model is established by using CPA, and the echo signal covariance matrix is calculated according to the position of physical array element. Then, after real value processing, the accurate low elevation is obtained by using maximum likelihood (ML) estimation or generalized multiple signal classification (MUSIC) algorithm. Finally, the target height is obtained by using geometric relationship. The simulation experiments compare the low elevation estimation performance of ULA and CPA virtual array, and prove the accuracy and superiority of the proposed low elevation estimation method based on physical array of CPA.