A comprehensive understanding of lake circulation is fundamental to inform better management of ecological issues and fishery resources in the Great Lakes. In this study, a high‐resolution, wave‐current coupled, three‐dimensional modeling system was applied to investigate the monthly and episodic dynamics of summer circulation in Lake Michigan. Model sensitivities to three wind sources and two grid resolutions against observed current velocities, water temperatures, and significant wave heights in the summer of 2014 were examined. Model performance was validated with additional satellite imageries and current measurements in the summer of 2015. Results indicated that the high‐resolution model driven by the observation‐based winds reproduced lake dynamics most reasonably. In July 2014, a pair of monthly averaged anticyclonic (i.e., clockwise) gyres in the surface layer were simulated in southern Lake Michigan. Analysis indicates that they originate from the wind‐driven, upwelling‐favorable, jet‐like Ekman currents along the west shore, which are connected by the density‐driven basin‐scale circulation. Although river inputs, strait exchanges, waves, grid resolutions, and bathymetric variations influence the monthly surface circulation, their effects are less important than the wind and density‐driven currents. Additional simulations support the predominant impacts of wind and density‐driven currents on lake surface circulation during a strong wind event. Further investigations suggest that lake circulation varies from surface to bottom layers, and this knowledge is significant to the related ecological issues and fishery resources management. The numerical model configured to Lake Michigan is beneficial to understanding dynamics in the Great Lakes system and other large water bodies.