Aiming at the difficulty of vibration signals feature extraction and condition identification of motor bearings，a method based on enhanced symplectic geometry mode decomposition（ESGMD）with self-organizing auto-encoder convolution network（SOAECN）is proposed. On the basis of symplectic geometry mode decomposition（SGMD），the vibration signals of motor bearing are adaptively decomposed into initial symplectic geometric mode component（ISGMCs）. The ISGMCs are adaptively reorganized by improved condensed clustering method to obtain the clustering symplectic geometric modal component （CSGMCs）. A new comprehensive evaluation index is proposed，which is used to screen and reconstruct the CSGMCs that can reflect the characteristics of vibration signals. Combined convolution neural network with wavelet auto-encoder，the auto-encoder convolution network（AECN）is constructed. Its loss function is improved and self-organizing strategy is introduced on the basis of AECN，then the SOAECN is constructed. The reconstructed vibration signals are fed into SOAECN for automatic feature learning and condition identification. Experimental results indicate that ESGMD-SOAECN reaches 98.98% of condition identification rate. The ability of condition automatic feature extraction and automatic condition identification is better than deep learning methods such as deep sparse auto-encoder，deep de-noising auto-encoder，deep belief network and so on. The results can provide a reference for the identification of motor bearing conditions.