<p>An adaptive compensation method with decoupled parameter identification is proposed to improve the disturbance rejection and stabilization precision of Inertially stabilized platforms (ISPs). We propose the confidence-weighted parameter identification method to model and decouple the nonlinear friction and mass imbalance torque by constant-speed rotation experiments. Furthermore, the adaptive compensation are introduced to reduce the oscillation and over-compensation for disturbance rejection. Then, we design a nonlinear sliding mode controller with adaptive laws to ensure asymptotic stability of the ISP under bounded uncertainties. The shock experiments verify that the proposed control system with parameter identification and adaptive compensation can effectively improve the dynamic tracking performance of the ISP. And, in the fire direction alignment experiments and the frequency sweep experiments, the proposed method enhances the anti-disturbance ability of ISPs. The control system with parameter identification and adaptive compensation for nonlinear friction and mass imbalance are of great significance in practical applications, which has a promising step towards improving the stability and accuracy of the ISP.</p>