Various proposed Mars missions, such as the Mars Sample Return Mission (MRSR) and the Mars Smart Lander (MSL), require precise re-entry terminal position and velocity states. This is to achieve mission objectives including rendezvous with a previous landed mission, or reaching a particular geographic landmark. The current state of the art footprint is in the magnitude of kilometers. For this research a Mars Precision Landing is achieved with a landed footprint of no more than 100 meters, for a set of initial entry conditions representing worst guess dispersions. Obstacles to reducing the landed footprint include trajectory dispersions due to initial atmospheric entry conditions (entry angle, parachute deployment height, etc.), environment (wind, atmospheric density, etc.), parachute deployment dynamics, unavoidable injection error (propagated error from launch on), etc. Weather and atmospheric models have been developed. Three descent scenarios have been examined. First, terminal re-entry is achieved via a ballistic parachute with concurrent thrusting events while on the parachute, followed by a gravity turn. Second, terminal re-entry is achieved via a ballistic parachute followed by gravity turn to hover and then thrust vector to desired location. Third, a guided parafoil approach followed by vectored thrusting to reach terminal velocity is examined. The guided parafoil is determined to be the best architecture. The purpose of this study is to examine the feasibility of using a computational intelligence strategy to facilitate precision planetary re-entry, specifically to take an approach that is somewhat more intuitive and less rigid, and see where it leads. The test problems used for all research are variations on proposed Mars landing mission scenarios developed by NASA. A relatively recent method of evolutionary computation is Particle Swarm Optimization (PSO), which can be considered to be in the same general class as Genetic Algorithms. An improvement over the regular PSO algorithm, allowing tracking of nonstationary error functions is detailed. Continued refinement of PSO in the larger research community comes from attempts to understand human-human social interaction as well as analysis of the emergent behavior. Using PSO and the parafoil scenario, optimized reference trajectories are created for an initial condition set of 76 states, representing the convex hull of 2001 states from an early Monte Carlo analysis. The controls are a set series of bank angles followed by a set series of 3DOF thrust vectoring. The reference trajectories are used to train an Artificial Neural Network Reference Trajectory Generator (ANNTraG), with the (marginal) ability to generalize a trajectory from initial conditions it has never been presented. The controls here allow continuous change in bank angle as well as thrust vector. The optimized reference trajectories represent the best achievable trajectory given the initial condition. Steps toward a closed loop neural controller with online learning updates are examined. The inner loop of the simulation employs the Program to Optimize Simulated Trajectories (POST) as the basic model, containing baseline dynamics and state generation. This is controlled from a MATLAB shell that directs the optimization, learning, and control strategy. Using mainly bank angle guidance coupled with CI strategies, the set of achievable reference trajectories are shown to be 88% under 10 meters, a significant improvement in the state of the art. Further, the automatic real-time generation of realistic reference trajectories in the presence of unknown initial conditions is shown to have promise. The closed loop CI...