Objectives. To provide baseline cohort descriptives and assess change in health behaviours since the UK COVID-19 lockdown.Design. A prospective cohort (N = 1,044) of people recruited online, purposively targeting vulnerable populations.Methods. After a baseline survey (April 2020), participants completed 3 months of daily ecological momentary assessments (EMA). Dietary, physical activity, alcohol, smoking, vaping and substance use behaviours collected retrospectively for the pre-COVID-19 period were compared with daily EMA surveys over the first 30 days during early lockdown. Predictors of behaviour change were assessed using multivariable regression models.Results. 30% of the cohort had a COVID-19 at risk health condition, 37% were classed as deprived and 6% self-reported a mental health condition. Relative to pre-pandemic levels, participants ate almost one portion of fruit and vegetables less per day (vegetables mean difference À0.33, 95% CI À0.40, À0.25; fruit À0.57, 95% CI À0.64, À0.50), but showed no change in high sugar portions per day (À0.03, 95% CI À0.12, 0.06). Participants spent half a day less per week doing ≥30 min of moderate to vigorous physical activity (À0.57, 95% CI À0.73, À0.40) but slightly increased days of strength training (0.21, 95% CI 0.09, 0.34), increased alcohol intake (AUDIT-C score change 0.25, 95% CI 0.13, 0.37), though did not change smoking, vaping or substance use behaviour. Worsening health behaviour change was associated with being younger, female and higher body mass index.Conclusions. The cohort reported worsening health behaviours during early lockdown. Longer term changes will be investigated using further waves of data collection.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.