The invention discloses an optimization control method based on self-adaptive virtual inertia parameters. The method comprises the following steps of: 1, acquiring a frequency change rate and a frequency offset of a grid-connected point of a system, and obtaining a virtual inertia coefficient by combining values of an anti-misoperation coefficient and a symbol item according to a corresponding criterion; 2, superposing an additional active instruction with a real-time power instruction issued by a fan master controller to obtain a total active instruction; 3, inputting the deviation value of the total active instruction and the fan active detection value into a PI regulator to obtain active and reactive current instructions; and 4, comparing the active and reactive current instructions with the actual active and reactive current, and respectively inputting the obtained current deviations into the PI regulator to obtain active and reactive voltage instructions, so that modulation is carried out by an SVPWM link, and a machine-side converter switching tube control signal is generated. According to the invention, the virtual inertia coefficient of the system can be smoothly adjusted according to the frequency fluctuation of the grid-connected point, thereby reducing the overshoot and adjustment time of the frequency response, and optimizing the inertia support capability of the system.