As SIFT algorithm has large calculation and the detection of Harris algorithm is unstable,the electronic image stabilization algorithm based on feature points matching is proposed.The feature points were extracted by Harris operator,and the extracted feature points were described by SIFT algorithm.Then the balance point of the computational complexity and matching accuracy for the algorithm was found.RANSAC criterion was used to improve the matching accuracy in the process of the feature points matching.The simulation results show that this algorithm can stabilize the jitter of the infrared video.