Background: Group CBT (G-CBT) for heterogeneous acute psychiatric inpatients (HAPIs), which allows patients to choose the group therapeutic target, might have clinical utility but is empirically untested. Aims: To test the feasibility, acceptability and patient-rated effectiveness of G-CBT for HAPIs in which patients' themselves choose the group therapeutic targets, within a previously rarely used sole-standalone session format. Method: Weekly G-CBT was run for two HAPI wards. The G-CBT was evaluated in terms of attendances/re-attendances, and patient feedback on 5-point scales of how strongly patients agreed/disagreed that the group was useful, enjoyable, worth re-attending, and had led to them learning something they could use to reduce their distress. Results: One hundred and thirty-seven separate patients attended a total of 291 times across 31 groups. Being female or having a diagnosis of bipolar disorder significantly predicted re-attendance. Sixty-three percent of patient feedback questionnaires were obtained from groups 10–31 and over 75% of respondents agreed positively with each of the evaluation dimensions. Conclusions: Practise-based evidence from this pilot study suggests that G-CBT for HAPIs, allowing patients to choose therapeutic targets in a sole-session format, is feasible, acceptable and patients find it effective. This supports more widespread deployment of this CBT treatment format. Future research might now test the format's clinical effectiveness with standardized and objective clinical outcome measures.