COVID-19 pandemic restrictions impacted dietary habits during the initial months of the pandemic, but long-term effects are unclear. In this longitudinal study, self-selected UK adults (n=1,733) completed three online surveys (May-June, August-September and November-December 2020, with a retrospective pre-pandemic component in the baseline survey), self-reporting sociodemographics, lifestyle and behaviours, including high fat, salt and sugar (HFSS) snacks, HFSS meals and fruit and vegetable (FV) intake. Data were analysed using generalised estimating equations. Monthly HFSS snacks portion intake increased from pre-pandemic levels (48.3) in May-June (57.6, p<0.001), decreased in August-September (43.7, p<0.001), before increasing back to pre-pandemic levels in November-December (49.2, p<0.001). 48.5% self-reported increased (25.9 [95% confidence interval: 24.1, 27.8]) and 47.7% self-reported decreased (24.1 [22.4,26.0]) monthly HFSS snacks portion intakes in November-December compared with pre-pandemic levels. Monthly HFSS meals portion intake decreased from pre-pandemic levels (7.1) in May-June (5.9, p<0.001), being maintained in August-September (5.9, p=0.897), and then increasing again in November-December (6.6, p<0.001), to intakes that remained lower than pre-pandemic levels (p=0.007). 35.2% self-reported increased (4.8 [4.3, 5.3]) and 44.5% self-reported decreased (5.1 [4.6,5.6]) monthly HFSS meals portion intakes in November-December compared with pre-pandemic levels. The proportion meeting FV intake recommendations was stable from pre-pandemic through to August-September (70%), but decreased in November-December 2020 (67%, p=0.034). Increased monthly HFSS snacks intake was associated with female gender, lower quality of life, and - in a time-varying manner - older age and higher HFSS meals intake. Increased monthly HFSS meals intake was associated with female gender, living with adults only and higher HFSS snacks intake. Reduced FV intake was associated with higher body mass index (BMI) and lower physical activity. These results suggest large interindividual variability in dietary change during the first year of the pandemic, with important public health implications in individuals experiencing persistent increases in unhealthy diet choices, associated with BMI, gender, quality of life, living conditions, physical activity and other dietary behaviours.