The current article deals with finite element (FE)- and genetic algorithm (GA)-based vibration energy harvesting from a tapered piezolaminated cantilever beam. Euler–Bernoulli beam theory is used for modeling the various cross sections of the beam. The governing equation of motion is derived by using the Hamilton's principle. Two noded beam elements with two degrees of freedom at each node have been considered in order to solve the governing equation. The effect of structural damping has also been incorporated in the FE model. An electric interface is assumed to be connected to measure the voltage and output power in piezoelectric patch due to charge accumulation caused by vibration. The effects of taper (both in the width and height directions) on output power for three cases of shape variation (such as linear, parabolic and cubic) along with frequency and voltage are analyzed. A real-coded genetic algorithm-based constrained (such as ultimate stress and breakdown voltage) optimization technique has been formulated to determine the best possible design variables for optimal harvesting power. A comparative study is also carried out for output power by varying the cross section of the beam, and genetic algorithm-based optimization scheme shows the better results than that of available conventional trial and error methods.