The combination of spatial information services, wireless mobile services, and social network services has generated massive social media geographic data, enriching the types of covert communication carriers. However, the imperceptibility and robustness of steganography methods based on geographic data still need to be improved. To this end, this paper proposes a steganography method, which hides secret messages into geo-coordinates. In this method, the sender analyzes the correlation between digit planes of geo-coordinate and positioning accuracy and divides digit planes into A-grade, B-grade, and C-grade according to significance. Then, secret messages are priority embedded in C-grade using the MLSBR algorithm to increase embedding capacity, and the rest are embedded into B-grade using STC adaptive steganography algorithm to enhance the imperceptibility. The receiver uses shared parameters to achieve the correct extraction of secret messages, and the extraction steps are obtaining stego geo-coordinates, dividing digit planes, and matching the steganography algorithm. Experimental results demonstrate that compared with the existing geo-coordinate steganography methods, the proposed method can realize the complete extraction of secret messages, the average embedding capacity is increased by 20.53 bits per geo-coordinate, the extraction error rate is reduced by 50.65%, the trajectory similarity is improved by 94.29%, and the PSNR is raised by 15.55 dB, which all show that the proposed method performs better under the comprehensive measure of average embedding capacity, imperceptibility, and robustness.