Background: Exercise delivered in the inpatient mental health setting may offer significant therapeutic benefit. Previous research has shown that acute group exercise at a self-selected intensity improves well-being, vigour, mood and calmness in people with mental illness. However, the impact of exercise undertaken at self-selected intensity on other feeling states is less well established. Any differences in the feeling state responses between consumers with differing diagnoses are important in individualising the delivery of exercise, maximising the potential benefits of exercise, and minimising adverse effects in the inpatient setting. Therefore, the purpose of the present study was to examine changes in feeling states in inpatient mental health consumers diagnosed with anxiety, bipolar or depressive disorders, following participation in a group exercise program performed at self-selected intensity. Methods: Forty adult (45.1 ± 13.8 years, 33.1 ± 9.6 kg m−2) inpatient mental health consumers with anxiety disorders (n = 14), bipolar disorders (n = 12) or depressive disorders (n = 14) completed the Exercise Induced Feeling Inventory prior to, and immediately following, an inpatient group exercise class comprising aerobic and resistance exercise performed at self-selected intensity. Within-person, paired sample t tests examined pre-post changes in Positive Engagement, Revitalisation, Tranquillity, and Physical Exhaustion. Results: For the group as a whole, participation in group exercise at self-selected intensity resulted in significant improvements in Positive Engagement (t(39) = 4.079, p < 0.001), Revitalisation (t(39) = 2.881, p = 0.006) and Tranquility (t(39) = 5.406, p < 0.001). When categorised by diagnosis, consumers with anxiety disorders showed a significant improvement in Tranquility (t(13) = 3.225, p = 0.007), while consumers with bipolar disorders showed significant improvement in Positive Engagement (t(11) = 2.538, p = 0.028), Revitalisation (t(11) = 3.125, p = 0.010), and Tranquility (t(11) = 4.690, p = 0.001). Consumers with depressive disorders showed a significant improvement in Positive Engagement (t(13) = 2.511, p = 0.026) only. Neither the group as a whole, nor any diagnostic group, showed an increase in Physical Exhaustion. Discussion: This study demonstrates that group exercise which includes aerobic and resistance exercise and is delivered in the impatient mental health setting improves feelings of Positive Engagement, Revitalisation, Tranquility, without significantly increasing feelings of Physical Exhaustion. These findings support the potential therapeutic role of group exercise delivered in the inpatient mental health setting. Important to consider, however, are the condition-specific differences in responses between consumers with anxiety, bipolar, or depressive disorders.