In the present work, numerical heat transfer analysis of fins with elliptic perforations is performed and a soft computing-based optimization method is proposed for its design. The computational analysis is performed by solving coupled heat and flow transport equations after validating the numerical model with an existing experimental result. A parametric study is done based on the numerical model to examine the effects of fin spacing, fin height, fin perforation major axis, minor axis lengths and the number of perforations. The analyzed model results in an increased heat transfer rate with volume reduction of up to 72% when compared to a rectangular solid fin. But the best heat transfer rate and volume reduction are not achieved for the same set of fin parameters. A multiple response optimization technique needed to arrive at a fin configuration that provides considerable reduction in weight along with an increase in heat transfer is proposed. It is difficult to obtain reasonable approximations using mathematical regressions for such multiple response optimizations as the system response is highly nonlinear. An attempt is made to use multiple output adaptive neuro-fuzzy inference system coupled with genetic algorithm to optimize the parameters of the fins with elliptical perforation considering increased heat transfer and weight reduction. The proposed new optimization algorithm is found to be more effective in determining the optimal parameters when compared to existing regression and soft computing methods for optimization.