To address the real-time optimization problem of the gold cyanidation leaching process, a self-optimizing control strategy is proposed. Based on the parametric uncertain model of leaching process, linear combinations of output measurements are selected as the controlled variables, which are tracked at constant set-points by feedback controllers to achieve the self-optimizing control of uncertain leaching process. Simulation results show the effectiveness of the proposed scheme. When disturbances occur, the system's economic performance is improved along with regulatory control and finally operated near the true optimum, hence more economic profit is gained for the leaching process.