Introduction: There is growing support for nature-based activities as potential effective interventions for depression. Randomized control trials (RCTs) to determine outdoor swimming benefits are needed. To inform the design of a full-scale trial, this feasibility study examines (1) participant recruitment, intervention and study engagement, (2) intervention and study safety, (3) participant experience and acceptability, (4) between-group mean difference estimates on measures of depression (primary clinical outcome), anxiety, wellbeing, mindfulness, self-compassion, health related quality of life and clinical resource use. Methods: Participants with mild to moderate symptoms of depression were recruited and randomised to an 8-session outdoor swimming course plus usual care or usual care only. Swimming courses took place in three locations. Participants completed online or paper surveys at baseline (T0), immediately after the intervention (T1, 8 weeks post-randomisation) and follow up 8 weeks later (T2, 16 weeks post-randomisation). Serious adverse events were recorded. Qualitative data were collected to explore participant, coach and social prescriber experiences. Results: 87 participants (99 % target) were recruited. 79 % (95 % CI: 61 %–88 %) of the outdoor swimming arm completed the intervention (4+ sessions out of 8). Two unrelated serious adverse events occurred during the trial. Participants' experience was positive and identified five themes for future study design: Accessibility, Belonging, Facing challenges with support, Benefiting and enjoyment and Clarity of information. Overall data completeness at all time points (including withdrawals and drop-outs) was high (85 %). There were between-group differences in favour of the intervention arm with medium to large effects across all measures. Conclusions: It is feasible to conduct a large-scale RCT to determine if an outdoor swimming course (in addition to usual care) can reduce symptoms of depression and is good value for money when compared to usual care alone.