Abstract: To mitigate the propelling cost of the six-strut tensegrity robot (TR-6), a BAGA (beetle antennae-genetic algorithm) hybrid optimization policy based on cable actuation is proposed. Firstly, the robot equivalent model is established via its initial geometrical configuration. A propelling optimization model is formulated, with the discrepancy of strain energy before and after actuation during each step as the objective function, which comprehensively considers constraints such as gravity moment, rolling energy, and cable regulation. Non-rigid-body motion analysis (NRMA) method is then employed to determine the robot posture in unbalanced status. The optimal robot propelling strategies are obtained utilizing BAGA, which are verified through ADAMS (automatic dynamic analysis of mechanical system) simulations. Finally, the effectiveness and practicality of the proposed method are confirmed through physical prototype testing of TR-6 with motor actuators. The results imply that the proposed method can effectively minimize the propelling cost of tensegrity rolling robots (single-cable scheme saves about 14% in strain energy compared to double-cable scheme), and provide a theoretical basis and technical support for the research of such robots.