OBJECTIVE Post-amputation rehabilitation is physically and cognitively demanding. Understanding which specific cognitive domains mediate outcome is critical to the development of interventions. METHODS A cohort undergoing post-amputation rehabilitation was assessed before limb fitting and followed up at 6 months (n = 34). The average age was 60.69 years (SD = 13.98). 82.4% of the sample was male. 79.4% had amputations because of peripheral arterial disease. Memory, visuospatial function, executive function, praxis, emotion and language were assessed at Time 1 (first prosthetic clinic attendance). Time 1 data were also gathered on aetiology, level of amputation, comorbidities, pain and demographics. Six month outcomes were the locomotor capability index (LCI), the special interest group in amputee medicine (SIGAM) mobility grades and self reported hours of use. RESULTS The LCI at 6 months was significantly predicted in regression analyses by a measure of visual memory (figure recall) (adjusted R2 = 24.8%, df = 32, zbeta = 0.52, p = 0.002. Hours of use were predicted by the verbal fluency test total (adjusted R2 = 17.1%, df= 26, zbeta = 0.45, p = 0.017). SIGAM mobility grades were predicted by a combination of immediate verbal memory (story recall), age, level of amputation and presence of pain (adjusted R2 = 58.2, df = 30, zbeta = 0.52, p = 0.000). CONCLUSIONS Neuropsychological and clinical variables predict a large amount of 6 month outcome variance. Cognitive difficulties may be considered mediators of poor outcome.