Optimal control problems for mechanical systems manifoldly arise in the fields of engineering and natural sciences, for instance, in robotics, biomechanics, automotive systems, or in space mission design. Here, the aim is to influence the systems dynamical behavior via its control inputs such that a given problem is optimally solved. For complex technical systems, an adequate modeling leads to hybrid dynamical systems. Hybrid control strategies provide a range of new possibilities for the design and the numerical computation of optimal control laws and, at the same time, a number of interesting open questions concerning the analysis, control, and optimization of hybrid dynamical systems arise. This thesis is devoted to two main aspects in this field: structure exploiting motion planning and optimal control of hybrid mechanical systems. The first part focuses on an optimal control method termed motion planning with motion primitives, which is based on exploiting inherent dynamical structures of the system, namely symmetry and (un)stable invariant manifolds. In many applications, energy efficiency plays a major role in the control design. To compute sequences with minimal control effort, the uncontrolled, i.e. natural dynamics are searched for (un)stable invariant manifolds of equilibrium points. In the second part of this thesis, the optimal control method DMOC (Discrete Mechanics and Optimal Control) is extended to hybrid mechanical systems. To this aim, a hybrid variational principle is developed. Then, an optimal control problem for a hybrid mechanical system can be addressed by a two layer approach. As an important subproblem, switching time optimization for discretized dynamical systems is studied. Finally, the motion planning method is brought into the hybrid optimal control setting.