The satellite-ground temperature difference was derived from the cloud top temperature of FY-4B and the air temperature of automatic weather stations. Combined with the cloud top height (CTH) observed by cloud radar, the satellite-ground vertical mean temperature lapse rate (SG-VMTR) was calculated and compared with the vertical temperature lapse rate (γ) fitted from the radiosonde temperature profile at the same stations. Meanwhile, the SG-VMTR was applied to other adjacent satellite pixels to achieve the fusion inversion of the satellite CTH in the surrounding areas of the cloud radar, and the inversion results were verified and analyzed. The results show that the SG-VMTR and γ values calculated at the same station were closely aligned and exhibited consistent temporal trends. The variation trends and statistical characteristics of the SG-VMTR at adjacent cloud radar stations were similar and can be used as substitutes to a certain extent. The correlation coefficient between the fused satellite-derived CTH and the cloud radar observed CTH was 0.90, while the root mean square error decreased by 720 m and the bias reduced to -17.7 m compared to pre-fusion results. This study lays a foundation for developing subsequent satellite-ground fusion regional CTH products.