This study addresses the optimization of lifting body shape variables with the goal of achieving optimal trajectory planning outcomes. Direct utilization of high-fidelity aerodynamic data for optimization is computationally expensive. To mitigate this, an 11-dimensional aerodynamic MFNN fusion model was developed, incorporating both shape variables and flight conditions to efficiently and accurately supply aerodynamic data for diverse configurations and operational scenarios. During the MFNN modeling phase, redundant features with negligible impact on modeling accuracy were identified and removed, thereby reducing input variable dimensionality and enhancing fusion modeling precision without increasing the scale of high-fidelity samples. Optimization was performed with the objective of maximizing range, constrained by factors such as minimum volume and maximum stagnation point heat flux. The optimized configuration demonstrated a 39.8% improvement in range. This research offers a viable approach for tackling multidisciplinary optimization challenges in aircraft design involving high-dimensional input variables.