Side match vector quantization method collects the possible codewords by using partial pixels of neighboring blocks. Furthermore, the state codebook design provides a way to improve compression ratio. In the complex content image, side match vector quantization method has a lower compression performance. The proposed method utilizes the property of discrete wavelet transform to save the computation cost for finding a suitable codeword. Moreover, a well designed novel state codebook by observing the characteristics of images in order to improve the compression ratio. In the test of smooth images, the visual quality of decompressed image by using our proposed method has approximated to the full search vector quantization. Moreover, the compression ratio for complex content images has a better performance than full search vector quantization and side match vector quantization.