An exact algorithm to compute an optimal 3D oriented bounding box was published in 1985 by Joseph O'Rourke, but it is slow and extremely hard to implement. In this article we propose a new approach, where the computation of the minimal-volume OBB is formulated as an unconstrained optimization problem on the rotation group SO (3,ℝ). It is solved using a hybrid method combining the genetic and Nelder-Mead algorithms. This method is analyzed and then compared to the current state-of-the-art techniques. It is shown to be either faster or more reliable for any accuracy.