Robotic fish design is an upcoming and interesting research area with lot of challenging tasks due to the impulsive dynamics of water space. In this paper an evolutionary computational approach is performed to design caudal fins under carangi form and sub-carangi form swimming modes. Size and Shape with SOLEIL and multi-body evolutionary experiments were carried out using Euler-Lagrangian equation-based BhT tool to experiment and validate the hydrodynamic effects of caudal fin by avoiding complex and time consuming CFD simulations to achieve realistic motion. To improve average velocity of robotic fish two approaches have been suggested, one is a hill climbing algorithm to find optimal shape with standard stiffness whereas the second approach considers both shape and stiffness together in a genetic algorithm. Finally simulated fin models are compared against physical models to identify the correlation and performances of both to accurately approximate real world performances in a simulated environment leading to design optimised caudal fins for robotic fish.