The cold bias in sea surface temperature (SST) over the Tropical Indian Ocean (TIO) in Climate Forecast System version 2 (CFSv2) has been a limiting hurdle in predicting the Indian summer monsoon rainfall (ISMR). The present study addresses the impact of increasing horizontal resolution of the atmospheric model of CFSv2 from a low resolution (T126, ~100 km) to a high resolution (T382, ~38 km) in improving the SST and Ocean mixed layer simulation. The high resolution run has reasonably reduced the SST cold bias (−0.8 to −0.2°C) over southern TIO and (−0.69 to 0.07°C) over northern TIO, however, a moderate warm bias of 0.37°C is observed over the tropical Pacific Ocean in T382 run. We have conducted a mixed layer heat budget analysis to understand the causative factors responsible for improved SST simulation. The improvement in net heat fluxes resulted in reduced biases in SST over TIO in T382 simulation. Better simulation of the entrainment process has further improved SST over Southern TIO. The high‐resolution atmospheric model is able to resolve convection centres reasonably well, therefore resulted in better simulation of winds, which enabled better prediction of Ocean mixed layer. The low and middle clouds are also better resolved in high resolution run, the total cloud cover is well predicted by the T382 run, hence the biases in radiative and heat fluxes are decreased significantly compared with its lower resolution model, which improved the radiation fluxes and responsible for the observed improvement in SST and mixed layer prediction. The dominance of biases in net shortwave radiation resulting in moderate positive SST biases over the Northern tropical Pacific Ocean.