Introduction: Qualitative studies suggest art participation in rehabilitation may improve post-stroke psychosocial outcomes. This study examined feasibility of conducting an RCT to evaluate such an art intervention.Method: Design: Feasibility randomised controlled trial. Population: Stroke survivors admitted for in-patient rehabilitation. Intervention Group (n = 41): 4–8 Creative Engagement Intervention (CEI) sessions to create a piece of artwork. Control Group: (n = 40) usual care and art portfolio to view. Outcomes: Stroke Impact Scale – social participation, communication, emotion; Positive and Negative Affect Schedule (PANAS); Visual Analogue Self-Esteem Scale (VASES), Trait Hope Scale; General Self-efficacy Scale (GSES), Self-Efficacy for Art (SEfA); Recovery Locus of Control. Blinded assessment: Baseline (T1), end of intervention (T2); 3 months (T3).Results: Of 198 admitted stroke survivors, 81 (41%) were recruited. 88% (n = 71) completed T2 and 77% (n = 62) completed T3 assessments. CEI mean change was greater than controls between T1–T2 and T1–T3 for Emotion, PANAS and SEfA. CEI mean change T1–T2 was also higher than controls for Social Participation. At T2, after adjusting for baseline covariates, estimated between group differences for SEfA (p < 0.001) and VASES (p = 0.01) significantly favoured intervention and control groups respectively. At T3 after adjusting for baseline covariates estimated between group difference for SEfA (p < 0.001) and GSES (p < 0.04) significantly favoured intervention and control groups respectively.Discussion: An RCT testing art participation in rehabilitation was possible and indicates art participation may benefit social participation, emotion, positive affect and self-efficacy for art after stroke. These potential effects warrant investigation in a fully powered RCT.