Entrance into college may be associated with elevated risk of progression in cigarette use among Chinese students. The purpose of this study was to examine patterns and dynamics of cigarette smoking by age and grade among college students in China. Cross-sectional data from 1845 students (44% females) in 19 colleges and universities in Jiangsu Province were derived from the Chinese Youth Health Risk Behavior Survey in 2001. Data were analyzed using latent class analysis (LCA). Overall 11 mutually exclusive types (latent classes) of cigarette users were derived and their proportions estimated from four variables: lifetime smoking, past 30-day smoking, attempting to quit in the past, and intention to smoke in the future. Two types (resistant never-smokers, self-stopped resistant nonsmokers) were observed across age groups and grades for both males and females while other smoking types varied with gender, age and grade. Advanced cigarette smoking types appeared in older age groups and higher grades. The cross-sectional age and grade trend suggests substantial progression in cigarette smoking among Chinese college students, underscoring the need for prevention among these students. Results also suggest that LCA may be an effective methodology for analyzing behavioral types and their dynamics with a limited number of variables.