In view of the disadvantage that the traditional Huffman coding algorithm needs to build a huffman tree. this paper presents a huffman coding algorithm that doesn' t rely on establishment of the huffman tree. The algorithm can put aside a specific tree structure, and obtain each symbol coding length from simulating tree creation process used only a one-dimensional array. At last, the algorithm can assign a code for each symbol according to the length of code. The example shows that the algorithm requires much less memory space than the traditional huffman coding algorith...