This communication proposes a method for designing a model-based onboard path planning unit for reusable launch vehicles. Flatness approach is used to map the system dynamics into a lower dimension space. As a consequence, the number of optimization variables involved in the optimal control problem is minimized. In addition, nonconvex nonlinear trajectory constraints in the flat output space are inner approximated by means of superquadric shapes. Genetic algorithms are used to find a global solution both for the superquadric shapes and the associated geometric transformations tuning parameters. Finally, simulations results are presented to illustrate the proposed approach. Simulations are based on the terminal area energy management phase of the Shuttle orbiter STS-1 reentry mission.