In this paper the optimization of a Tubular Permanent Magnet-Linear Generator (TPM-LiG) for energy generation is presented. The application is related to the sea wave energy generation for small sensorized buoy. The optimization process is developed by means of an hybrid evolutionary algorithm widely presented in the paper. The advantage of this algorithm is in the wide exploration of the variables space and in the effective exploitation of the fitness function. The algorithm has been tested on a benchmark case and then applied to the optimization of the here considered device.