Objective The objective of this study was to explore profiles of quality of life (QoL) trajectories during the natural history of dementia and individual variations contributing to QoL trajectories. Methods We conducted a longitudinal community‐based study of 520 elderly people with mild cognitive impairment and 100 healthy people aged 60 years or over. We conducted six waves of assessment between October 2010 and May 2013 in Taiyuan, mainland China. Cognitively normal, mild cognitive impairment, global impairment, and Alzheimer's disease (AD) were defined as state 1, 2, 3, and 4, respectively. We assessed health‐related QoL (HRQoL) via the Quality of Life–Alzheimer's Disease (QoL–AD) Chinese version. We used the latent growth curve model (LGCM) to investigate change in HRQoL over time. Results Latent growth curve model analysis revealed a mean initial QoL level of 29.865 with substantial variation and a significant mean slope for the whole sample. Multigroup LGCM showed substantial variations across individuals in initial QoL levels for each cognitive state transition group. For the slope factor, we found significant changes and variations for the transition from state 2 to 3 and from state 3 to 4. We estimated mean QoL levels over six assessments based on intercept, slope, and factor loadings for the whole sample and the three cognitive state transition groups. Conclusions A decline in subjective QoL is not inevitable during the natural history of dementia in community settings, and there is a degree of individual variation in QoL. Future studies should investigate the factors associated with individual variations in QoL trajectories in AD.