Middle spatial resolution multi-spectral remote sensing image is a kind of color image with low contrast, fuzzy boundaries and informative features. In view of these features, the fuzzy C-means clustering algorithm is an ideal choice for image segmentation. However, fuzzy C-means clustering algorithm requires a pre-specified number of clusters and costs large computation time, which is easy to fall into local optimal solution. In order to overcome these shortcomings, ant colony algorithm is employed to optimize fuzzy C-means algorithm in remote sensing image segmentation. First, the centers and number of clusters is determined by ant colony optimization algorithm. Then the initialization fuzzy C-means algorithm is used for remote sensing image classification. Experimental results show that the ant colony optimization is an effective method to solve the problem of fuzzy C-means algorithm in remote sensing image segmentation and the visual interpretation of segmentation is much improved by proposed ant colony optimized C-means clustering