The COVID-19 pandemic provided a unique opportunity to examine the contributions of social interaction and other non-social factors to loneliness among older adults in the context of confinement measures constraining opportunities for in-person social interactions. This study aims to identify groups of individuals with heterogeneous trajectories of loneliness during the COVID-19 pandemic and to explore the sociodemographic, health, psychological and social interaction-related factors associated with these trajectories. In this 12-month longitudinal study, 614 community-dwelling individuals aged 60+ years completed telephone-based interviews on four occasions between May 2020 and May 2021. Loneliness was evaluated using the three-item version of the UCLA Loneliness Scale. Multilevel modelling assessed average changes in loneliness over time. Group-based trajectory modelling was performed to identify distinct trajectories of loneliness over time. Multinomial logistic regressions were conducted to explore the predictors of these trajectories. On average, there was a curvilinear change in loneliness that tracked the stringency of the COVID-19-related confinement measures. In this convenience sample, three heterogeneous trajectories were identified: a stable-low (17.2%), a fluctuating-moderate (48.8%) and a sustained-elevated (34.0%) trajectory. Participants in the sustained-elevated loneliness trajectory were more likely to live alone and experience elevated psychological distress and greater COVID-19 perceived health threat compared to those in the stable-low trajectory. Participants in the fluctuating-moderate loneliness group were more likely to have multimorbidity, experience greater psychological distress, and have less frequent in-person interactions than the stable-low loneliness group. Assessing the combination of sociodemographic, health, psychological and social factors may help identify individuals at higher risk for chronic loneliness.