Abstract Background Therapeutic exercise is an effective intervention for knee pain and osteoarthritis (OA) and should be individualised. In a preliminary, proof-of-principle study we sought to develop a home exercise programme targeted at specific physical impairments of weak quadriceps, reduced knee flexion range of motion (ROM) and poor balance, and evaluate whether receipt of this was associated with improvements in those impairments and in patient-reported outcomes among older adults with knee pain. Methods This community-based study used a single group, before-after study design with 12-week follow-up. Participants were 58 adults aged over 56 years with knee pain and evidence of quadriceps weakness, loss of flexion ROM, or poor balance, recruited from an existing population-based, observational cohort. Participants received a 12-week home exercise programme, tailored to their physical impairments. The programme was led, monitored and progressed by a physiotherapist over six home visits, alternating with six telephone calls. Primary outcome measures were maximal isometric quadriceps strength, knee flexion ROM and timed single-leg standing balance, measured at baseline, 6 and 12 weeks by a research nurse blinded to the nature and content of participants’ exercise programmes. Secondary outcome measures included the WOMAC. Results At 12 weeks, participants receiving strengthening exercises demonstrated a statistically significant change in quadriceps isometric strength compared to participants not receiving strengthening exercises: 3.9 KgF (95 % CI 0.1, 7.8). Changes in knee flexion ROM (2.1° (−2.3, 6.5)) and single-leg balance time (−2.4 s (−4.5, 6.7)) after stretching and balance retraining exercises respectively, were not found to be statistically significant. There were significant improvements in mean WOMAC Pain and Physical Function scores: −2.2 (−3.1, −1.2) and −5.1 (−7.8, −2.5). Conclusions A 12-week impairment-targeted, home-based exercise programme for symptomatic knee OA appeared to be associated with modest improvements in self-reported pain and function but no strong evidence of greater improvement in the specific impairments targeted by each exercise package, with the possible exception of quadriceps strengthening. Trial registration Clinical Trial Registration Number: ISRCTN 61638364 Date of registration: 24 June 2010