The use of ADRC to control optoelectronic stabilization platforms has been widely used due to its strong resistance to external disturbances, good robustness, and no need for accurate disturbance modeling. Nonlinear ADRC has better anti-disturbance and control effects due to its nonlinear characteristics. However, the nonlinear function in nonlinear ADRC currently has room for improvement due to insufficient use of measurable information, a single function form that cannot be changed, and possible failure of nonlinear characteristics. To this end, this paper designs a nonlinear function with multiple inputs and variable forms, and uses an optimization algorithm to optimize the parameters.This paper proposes a fal function with multiple inputs and variable forms. The system input change signal and the system output error signal are added to the nonlinear function (Eq.5), and the function form is fitted using a polynomial (Eq.6). The convergence of the extended state observer is proved by using the sufficient conditions for asymptotic stability in the sense of Lyapunov. Due to the large number of adjustable parameters, the improved raccoon algorithm is used to optimize the control parameters, the population initialization distribution is optimized using the good point set (Eq.22), the global optimization is optimized using the logistic mapping (Eq.24), and finally the controller parameters are adjusted using the improved raccoon algorithm.The improved nonlinear function enhances the anti-disturbance capability of the system by increasing input variables and adjustable function forms. The simulation comparison of the improved algorithm shows that the MIAfal function reduces the errors caused by disturbances compared with the fal function (Fig.7). The test function and simulation comparison of the raccoon algorithm before and after the improvement are carried out. The test function shows that the improved raccoon algorithm has obvious advantages in optimization speed and accuracy (Fig.6). The simulation shows that the control parameters obtained by the improved optimization algorithm have stronger ability to suppress disturbances (Fig.8). The control ability of the MIAfal function was verified by experiment design using a certain type of stable platform (Fig.9) with a swinging platform. The results show that under signals with frequencies of 0.5 Hz, 1 Hz, 2 Hz, and 3 Hz and amplitudes of 0.17 rad/s, 0.1 rad/s, 0.05 rad/s, and 0.03 rad/s, respectively, the control system using the MIAfal function reduced the error by 33.1%, 41.2%, 37.4%, 29.6% (Fig.10) compared with the control system using the fal function, indicating that the MIAfal function has better disturbance suppression capability.The nonlinear function was improved by adding input variables and changing the function form, and the control parameters were optimized by the improved raccoon algorithm, which enhanced the anti-disturbance ability of the anti-disturbance control. Test functions and simulations show that the improved raccoon algorithm has faster convergence time and more accurate optimization accuracy. Simulations and experiments show that the MIAfal function has significantly stronger anti-disturbance ability than the fal function under speed input signals within 0.5-3 Hz, which can effectively improve the performance of the optoelectronic stabilization platform system.