Accurate computation of geometric moments is very important in computer vision, image processing and pattern recognition. In this paper, inspired by the idea of bit-plane decomposition and accurate computation of geometric moments on binary images, we put forward an accurate and fast algorithm for the computation of geometric moments using Non-symmetry and Anti-packing Model (NAM) for color images, which takes O(N) time where N is the number of all NAM blocks. By taking four color images 'Lena', 'Peppers', 'Frog', and 'Fish' as typical test objects, and by comparing our proposed NAM-based accurate algorithm with the popular Binary Tree (BT)-based accurate algorithm for computing the geometric moments, the theoretical and experimental results presented in this paper show that our NAM-based accurate algorithm can significantly improve the execution speed by 43.71%, 41.93%, 41.01%, and 38.63% over the BT-based accurate algorithm in images 'Lena', 'Peppers', 'Frog', and 'Fish', respectively. Also, our NAM-based accurate algorithm can significantly improve the average execution speed by 41.32% over the BT-based accurate algorithm. Therefore, in the case of computing lower order moments of color images, our proposed accurate algorithm is much faster than the BT-based accurate algorithm.