Chemical explosives are among the available high energy-dense storage materials with transferable energy to surroundings or adjacent materials during detonation. The effectiveness, ignition energy as well as spark sensitivity of these explosives are governed by the velocity of detonation which needs to be determined before energetic material synthesis purposely to enhance safety and lessen the cost as well as the difficulties associated with material synthesis and evaluation. This present research work proposes hybridization of support vector regression (SVR) and genetic algorithm (GA) for estimation of the velocities of primary explosives for the first time. The performance of the proposed hybrid SVR-GA is compared with the existing Mohammad Jafari et al. model as well as the results of EXPLO5Code's prediction using four parameters for measuring model's performance. The proposed SVR-GA model shows superior performance compared with two existing models in the literature. The performance of the proposed SVR-GA model for explosive velocity estimation strengthens its practical application, circumvents the experimental challenges and minimizes the associated potential risk.