Attitude measurement error is one of the main factors that deteriorate the imaging accuracy of airborne remote sensing. In view of the fact that high-precision inertial navigation instruments are very costly, a low-cost but effective method based on the aerial images is proposed to improve the accuracy of attitude measurement. In this method, a large number of aerial images are shot on the flight. The image matching method is used to calculate the relative movement of each two adjacent aerial images. Based on the relationship between the changes of the attitude angles at the two shooting time and the relative movement of the two adjacent aerial images, the changes of the attitude angles can be obtained, which can be used to rectify the erroneous attitude data measured by low-accuracy inertial navigation instruments. The realization of the method was specifically analyzed. The simulation results also demonstrate the effectiveness of the method.