High compression ratio,high decoding performance,and progressive data transmission are the most important require-ments of vector data compression algorithms for WebGIS.To meet these requirements,we present a new compression approach.This paper begins with the generation of multiscale data by converting float coordinates to integer coordinates.It is proved that the distance between the converted point and the original point on screen is within 2 pixels,and therefore,our approach is suitable for the visualization of vector data on the client side.Integer coordinates are passed to an Integer Wavelet Transformer,and the high-frequency coefficients produced by the transformer are encoded by Canonical Huffman codes.The experimental results on river data and road data demonstrate the effectiveness of the proposed approach:compression ratio can reach 10% for river data and 20% for road data,respectively.We conclude that more attention needs be paid to correlation between curves that contain a few points. High compression ratio, high decoding performance, and progressive data transmission are the most important require-ments of vector data compression algorithms for WebGIS. To meet these requirements, we present a new compression approach. This paper begins with the generation of multiscale data by converting float coordinates to integer coordinates.It is proved that the distance between the converted point and the original point on screen is within 2 pixels, and therefore, our approach is suitable for the visualization of vector data on the client side. Integer coordinates are passed to an Integer Wavelet Transformer, and the high-frequency coefficients produced by the transformer are encoded by Canonical Huffman codes. The experimental results on river data and road data demonstrate the effectiveness of the proposed approach: compression ratio can reach 10% for river data and 20 % for road data, respectively.We conclude that more attention needs be paid to correlation between curves that contain a few points