BACKGROUND: No previous studies have reported predictors and moderators of outcome of psychological therapies for depression experienced by adults with intellectual disabilities (IDs). We investigated baseline variables as outcome predictors and moderators based on a randomised controlled trial where behavioural activation was compared with guided self-help. METHODS: This study was an exploratory secondary data analysis of data collected during a randomised clinical trial. Participants (n = 161) were randomised to behavioural activation or guided self-help and followed up for 12 months. Pre-treatment variables were included if they have previously been shown to be associated with an increased risk of having depression in adults with IDs or have been reported as a potential predictor or moderator of outcome of treatment for depression with psychological therapies. The primary outcome measure, the Glasgow Depression Scale for Adults with Learning Disabilities (GDS-LD), was used as the dependant variable in mixed effects regression analyses testing for predictors and moderators of outcome, with baseline GDS-LD, treatment group, study centre and antidepressant use as fixed effects, and therapist as a random effect. RESULTS: Higher baseline anxiety (mean difference in outcome associated with a 1 point increase in anxiety 0.164, 95% confidence interval [CI] 0.031, 0.297; P = 0.016), lower performance intelligence quotient (IQ) (mean difference in outcome associated with a 1 point increase in IQ 0.145, 95% CI 0.009, 0.280; P = 0.037) and hearing impairment (mean difference 3.449, 95% CI 0.466, 6.432; P = 0.024) were predictors of poorer outcomes, whilst greater severity of depressive symptoms at baseline (mean difference in outcome associated with 1 point increase in depression -0.160, 95% CI -0.806, -0.414; P < 0.001), higher expectation of change (mean difference in outcome associated with a 1 point increase in expectation of change -1.013, 95% CI -1.711, -0.314; p 0.005) and greater percentage of therapy sessions attended (mean difference in outcome with 1 point increase in percentage of sessions attended -0.058, 95% CI -0.099, -0.016; P = 0.007) were predictors of more positive outcomes for treatment after adjusting for randomised group allocation. The final model included severity of depressive and anxiety symptoms, lower WASI performance IQ subscale, hearing impairment, higher expectation of change and percentage of therapy sessions attended and explained 35.3% of the variance in the total GDS-LD score at 12 months (R2  = 0.353, F4, 128  = 17.24, P < 0.001). There is no evidence that baseline variables had a moderating effect on outcome for treatment with behavioural activation or guided self-help. CONCLUSIONS: Our results suggest that baseline variables may be useful predictors of outcomes of psychological therapies for adults with IDs. Further research is required to examine the value of these potential predictors. However, our findings suggest that therapists consider how baseline variables may enable them to tailor their therapeutic approach when using psychological therapies to treat depression experienced by adults with IDs.