More evidence from prospective studies is needed to determine if and how social cognitive constructs mediate behaviour change. In a longitudinal study, we aimed to examine potential social cognitive mediators of objectively measured physical activity (PA) behaviour among people with type 2 diabetes (T2D) who participated in a six-month PA intervention. All participants from the proven effective Healthy Eating and Active Living for Diabetes in Primary Care Networks trial were included for this secondary analysis. Change in pedometer-derived daily step counts (baseline to six months) was the outcome of interest. Primary constructs of interest were from Social Cognitive Theory, however constructs from and Theory of Planned Behaviour were also tested in a mediating variable framework using a product-of-coefficients test. The sample (N=198) had a mean age of 59.5 (SD 8.3) years, haemoglobin A1c 6.8% (SD 1.1), 50% women, BMI 33.6kg/m(2) (SD 6.5), systolic pressure 125.6mmHg (SD 16.2) and average daily steps were 5879 (SD 3130). Daily pedometer-determined steps increased for the intervention group compared to usual care control at six-months (1481 [SD 2631] vs. 336 [SD 2712]; adjusted p=.002). There was a significant action theory test effect for planning (A=.21, SE=.10, p=.037), and significant conceptual theory test results for subjective norms (B=657, SE=312, p=.037) and cons (B=-664, SE=270, p=.015). None of the constructs satisfied the criteria for mediation. We were unable to account for the effect of a pedometer-based PA intervention for people with T2D through our examination of mediators. Our findings are inconsistent with some literature concerning PA interventions in diabetes; this may be due to variability in measures used or in study populations.