Cirrus clouds play an important role in atmospheric radiation and affect weather systems and climate change. Satellite remote sensing has considerable advantage in cirrus cloud detection, relative to traditional observation. As a passive remote sensing instrument, large deviations in the thin cirrus cloud top height data from the Moderate Resolution Imaging Spectroradiometer (MODIS) are detected. Comparatively, the Cloud–Aerosol Lidar with Orthogonal Polarization (CALIOP) aboard the Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO), which is an active remote sensing instrument, can acquire more accurate characteristics of thin cirrus cloud. In this study, MODIS cloud products in the Beijing–Tianjin–Hebei region from 2013 to 2017 are selected. Using the CALIPSO cirrus cloud top height data, a linear fitting method based on the cross-validation method is obtained, and the MODIS cirrus cloud top height data are corrected. The difference between MODIS and CALIPSO changes from -3 to 2 km to -2.0 to 2.5 km. Moreover, the maximum difference changes from approximately -0.8 km to approximately 0.2 km. In the context of different vertical levels and cloud optical depths, MODIS cirrus cloud top height data are improved after correction, which is more obvious at the lower cloud top height and optically thinner cirrus clouds.