<p>This study presents a computational framework for optimizing material orientations in geometrically nonlinear inflatable shell structures. The proposed method employs a tensor-field representation of material orientation, which avoids the nonlinear constraints and numerical instabilities often encountered in conventional angular or vector-based formulations. In particular, a relaxed orientation tensor, whose trace may take values less than unity, is introduced to allow smooth optimization and to represent a moderate reduction in material stiffness when required by the deformation state. The shell structures are modeled using the three-node Mixed Interpolation of Tensorial Components (MITC3) element with displacement-dependent follower forces to represent the internal pressure. The material is assumed to be transversely isotropic with a single preferred direction. The optimization problem is solved within a gradient-based framework incorporating Helmholtz filtering and a smoothed Heaviside projection to achieve smooth and physically meaningful orientation fields. Sensitivities of the objective function with respect to the design variables are efficiently computed using an adjoint method combined with the complex-step derivative approximation. Several numerical examples are provided to demonstrate the accuracy, robustness, and effectiveness of the proposed optimization framework.</p>