JAXA develops an unmanned technology demonstrator vehicle for low sonic boom research. The research project is called D-SEND #2 (Drop test for Simplified Evaluation of Non-symmetrically Distributed sonic boom). The autonomous demonstrator vehicle will be dropped from a high altitude balloon, and it will accelerate and glide to fly over a predetermined measurement area with a supersonic cruise condition. Trajectory optimization analysis is necessary for the design of onboard reference trajectory generation logic, which is robust against the dropping point uncertainty. An easy-to-use trajectory optimization tool using dynamic programming approach has been developed in order to generate reference trajectories in the early stage of design process. This paper introduces how the dynamic programming approach is applied to the D-SEND #2 flight trajectory optimization, and it shows numerical examples, which include the maximum and minimum range problems, and range adjustment problems.