Atomic force microscopes (AFMs) are usually utilized for nano-scale imaging. Usually, the slide loading the samples and the stage of the AFM cannot be placed completely parallel with the motion plane of the piezoelectric actuator due to the manual operation and machining errors, which leads to the variation of imaging brightness (or height) along the slope and makes the relative topography unrealistic. Considering the fact that the introduced slopes are different for every imaging process, this paper proposes an effective and efficient real-time preprocessing approach for slope elimination, where the recurrent least square method is firstly utilized to estimate the slope, and the imaging method is then combined together to eliminate the slope and display the image in a real-time manner. Both simulation and experimental results demonstrate the superior performance of the proposed method.