Two-stage exo-reversible thermoelectric generator considering internal irreversibilties is developed in electrically series and electrically parallel configurations in matrix laboratory simulink environment based on finite time thermodynamic analysis. Simultaneous optimization of proposed system for maximizing power output, thermal efficiency and minimizing entropy generation is employed based on second version of non-dominated sorting genetic algorithm, where working electrical current, number of thermoelectric elements in top and bottom stage, temperature of hot side and cold side are considered as design variables. The present work explores optimal values of above performance parameters and design variables with the comparison of two-stage thermoelectric generator in two modes. For triple and dual objective functions, Pareto frontier is obtained and using four decision variable techniques viz. Fuzzy Bellman-Zadeh, Shannon's entropy, LINMAP and TOPSIS best optimal solution is selected. With the current study, it has been proved that triple-objective optimization gives much lower difference between ideal and obtained solution, termed as deviation index, as compared to the dual objective optimization. (C) 2016 Elsevier Ltd. All rights reserved.