Owing to diversification assembly states in batch assembly， feature positioning errors， which severely interrupt the process and affect the efficiency， are easily caused. Therefore， establishing a strong robust feature localization algorithm is essential. This paper proposes a support vector machine model that synthesizes gradient histograms and local binary patterns. The pyramid search strategy is utilized to improve the recognition efficiency and develop a micro-feature localization method. Both performance verification and heuristic applications are performed on self-developed precision automatic assembly equipment， and different features are recorded for support vector machine training. The influences of interference factors， such as texture and illumination， on the positioning stability are investigated in detail. Additional experiments on the positioning accuracy and actuator component assembly are performed. The results reveal that the proposed approach exhibits good unimodal， repetitive accuracy and robustness under various conditions. The recognition accuracy rate can reach 98%. Its positioning accuracy is better than 4 μm， and the actual assembly accuracy is better than 7 μm. The feature localization method can meet localization requirements under different assembly conditions in actual batch production and provides an effective solution for precision automatic assembly localization.