Background: The long-term recovery (in years) of patellar tendinopathy treated with loading-based rehabilitation remains largely unknown. Purpose: To examine the clinical outcome and tendon structure years after exercise-based treatment of chronic patellar tendinopathy. Study Design: Case series; Level of evidence, 4. Methods: This was a 3- to 4-year follow-up evaluation of participants (N = 28) from a previous randomized clinical trial by the author group. All participants received loading-based rehabilitation for 12 weeks with either moderate-slow resistance (55% of 1 repetition maximum) or heavy-slow resistance (up to 90% of 1 repetition maximum). Both groups showed similar improvements after 3 and 12 months and were therefore collapsed in the present analysis. Function and symptoms (the Victorian Institute of Sport Assessment–Patella [VISA-P] questionnaire), tendon pain (numeric rating scale [NRS] during activity and during a single-leg decline squat test), and tendon structure (tendon vascularization and thickness on ultrasound) were assessed. Results: The mean follow-up was 3.6 ± 0.4 years after the baseline assessment in the original clinical trial. The VISA-P score was 83.9 ± 11.9 (95% CI, 79.3-88.5) at the latest follow-up and did not differ from the 1-year follow-up score ( P = .54). Similarly, NRS score during preferred sport (1.6 ± 1.7; 95% CI, 0.9-2.2) and single-leg decline squat (1.0 ± 1.8; 95% CI, 0.3-1.7) did not differ from the 1-year values and remained elevated. Power Doppler area and tendon thickness decreased significantly from 1 year to latest follow-up ( P < .0001 and P = .02, respectively), but power Doppler area >1 mm2 was still present in 43% of the participants after 3 to 4 years, and the tendon thickness was still mildly elevated (6.4 ± 1.8 mm; 95% CI, 5.7-7.1 mm). Sports participation after 3 to 4 years (3.9 ± 2.7 h/wk; 95% CI, 2.9-7.1 h/wk) was significantly lower compared with preinjury levels ( P < .0001). Conclusion: Clinical symptoms remained even years after loading-based treatment for patellar tendinopathy, whereas some but not all tendon structures normalized in this longer term follow-up.