Nowadays, armored vehicles have realized the comprehensive utilization of the whole vehicle energy through the power network. Load frequency control is the basic control approach to ensure the reliable operation of the armored vehicle power system (AVPS). An optimal load frequency control approach based on adaptive dynamic programming (ADP) and a disturbance observer is proposed for an AVPS with unknown disturbances. Owing to the existence of unknown disturbances in the AVPS, a nonlinear disturbance observer is designed to estimate unknown disturbances, and the disturbance compensation control is designed to address the influence. Then, a load frequency optimal control is designed under the framework of ADP to solve the optimal control problem of load frequency in the AVPS. It should be mentioned that the ADP method uses critic networks and action networks to approximate the performance index function and the optimal control signal, respectively. Finally, based on Lyapunov theory, we can prove that all states in an AVPS with disturbances are bounded, and the effectiveness of the proposed control approach is verified by simulation.