A coupled ocean–atmosphere–wave model was used to assess the impact of model coupling on the simulations of air–sea fluxes, surface currents, waves, and temperature profile during the passage of a tropical cyclone (TC) Phailin in the Bay of Bengal. Four numerical experiments with different coupling configurations among the atmosphere, ocean, and wave models were carried out to identify differences in simulated atmospheric and oceanic parameters. The simulated track and intensity of Phailin agree well with the observations. The inter-comparison of model experiments with different coupling options highlights the importance of better air–sea fluxes in the coupled model as compared to the uncoupled model towards an improvement in the simulation of TC Phailin. The coupled model configurations overcome the cold bias (up to − 2 °C) in sea surface temperature simulated by the uncoupled ocean model. A higher magnitude of the surface drag coefficient in the uncoupled atmosphere model enhanced the bottom stress (> 2 N m−2). As a result of excess momentum transfer to the sea surface, the uncoupled ocean model produced stronger surface currents as compared to the coupled model. The inclusion of the wave model increases the sea surface roughness and, thereby, improves the wind speed and momentum flux at the air–sea interface. The maximum significant wave height in the coupled model was about 2 m lower than the uncoupled wave model. The model experiments demonstrate that the periodic feedback among the atmosphere, ocean, and wave models leads to a better representation of momentum and heat fluxes that improves the prediction of a tropical cyclone.