The invention provides a binocular vision-based target attitude calculation method, which comprises the steps of (1) collecting multiple groups of images through left and right cameras and completing calibration of the left and right cameras; (2) carrying out correcting and filtering processing on the images collected by the left and right cameras; (3) extracting a target from the images obtained in the step (2), processing to obtain a target contour and a coordinate position of the target in the image; (4) dividing a picture comprising the target from the image in the step (2), extracting feature points in the divided picture and correcting coordinates of the feature points through the position information of the target obtained in the step (3) in the image; (5) selecting the optimal feature point according to the dispersion degree of the feature points; and (6) calculating the target attitude by using the optimal feature point. According to the method, the binocular vision-based feature point coordinate calculation accuracy can be improved, and the feature point matching precision can be improved and the robustness and the stability of an attitude calculation algorithm can be improved.