Objective Based on latent growth mixture model(LGMM)and China Health and Retirement Longitudinal Survey(CHARLS)Phase 3 tracking data,to investigate the development trajectory and influencing factors of the cognitive level of the elderly.Methods The data were from the data of CHARLS in 2013,2015,and 2018,including a total of 4662 elderly people aged 60 and above as research samples.The cognitive level of the elderly was assessed using the MMSE scale.Using the Mplus tool to construct an LGMM model suitable for cohort data to explore the trajectories of cognitive level changes in different elderly groups,examining time-constant factors(gender,age,education level,marital status,smoking,drinking)and time-varying factors(daily life activity ability,sleep duration,subjective life satisfaction)on the cognitive level of the elderly.Results According to the relevant indicators,the development trend of the cognitive level of the elderly can be divided into three types,namely the C1 slow decline group(35.2%),the C2 rapid decline group(20.6%)and the C3 stable group(44.2%).The cognitive level of people showed an overall downward trend with the passage of time.The cognitive level development trajectories of the three groups had significant time effects.The cognitive level of the C2 rapid decline group declined rapidly with time.The cognitive level of and the C1 slow decline group decreased slowly over time,and the cognitive level of the C3 stable group did not change significantly and was in a steady state.Conclusion The developmental trajectory of the cognitive level of the elderly has significant group heterogeneity and individual differences,and there are different types of development trends.