Based on the latest data of China Statistical Yearbook， this paper uses sparse principal component and cluster analysis methods to study the difference of urban construction level in various regions of China in 2017 and 2018. Sparse principal component analysis uses a penalty function method to sparse the loading matrix while reducing the dimensionality of data， so that the analysis results are more interpretable than principal component analysis. Moreover， cluster analysis is carried out for each region in China， and the results verify the rationality of the ranking of the sparse principal component method. The results show that the level of urban construction in China is uneven, the level of urban construction in Jiangsu， Zhejiang， Guangdong and Shandong is higher than Tibet， Ningxia and Guizhou. Finally， some suggestions are provided for the development of urbanization in all parts of China.