This paper investigates the optimal control problem for rigid spacecraft attitude maneuver in the presence of inertial uncertainties. A neural network identifier is designed to deal with the unknown dynamics of the spacecraft system. Based on the identification results, an optimal control law is proposed using the adaptive dynamic programming framework. To address the challenge of solving the Hamilton-Jacobi-Bellman equation, a critic neural network is constructed to obtain the nearly optimal control law. It is proven in the Lyapunov sense that the closed-loop system states and the weight estimation error are uniformly ultimately bounded. Numerical simulation results demonstrate the effectiveness of the proposed control scheme.