A methodology is presented to linearly correct the tendency of sea surface temperature (SST) anomalies in a coupled model. Using an atmospheric general circulation model (AGCM) coupled to a slab ocean as an example, we demonstrate the effectiveness of the linear correction methodology in improving the model's skill predicting SST in the tropical Atlantic Ocean during boreal spring. For this particular coupled model, the correction mainly takes into consideration the linear ocean dynamics absent in the slab ocean, thereby improving the skill in the tropical south and equatorial Atlantic. The corrected coupled model is further shown to produce a skillful rainfall forecast in the intertropical convergence zone (ITCZ) region during the boreal spring.