Performance analysis and optimization of a circular pin-fin with inside gaps in a rectangular cooling channel were performed at Reynolds number, 10,000, using three-dimensional Reynolds-averaged Navier-Stokes equations and a multi-objective genetic algorithm. The low-Reynolds-number version of the shear stress transport model was used as turbulence closure. A parametric study was also performed to identify the geometrical effects of the pin-fin on heat transfer and pressure drop. The straight and reference gapped pin-fins yielded better performances than those of the circular pin-fin without the gap in terms of both heat transfer and pressure drop. The objective of the optimization was to maximize the heat transfer and minimize the pressure loss, simultaneously. The area-averaged Nusselt number and pressure loss coefficient were considered as objective functions, and three design variables related to the geometry of the gapped pin-fin were chosen for the optimization. Twenty-seven design points were generated using Latin hypercube sampling in the design space, and response surface approximation models were constructed for the objective functions. The optimization results were analyzed using five representative solutions on the Pareto-optimal front. The objective functions were found to be significantly affected by variation in the design variables, especially, the width of front gap and the rear gap angle.