We reviewed the records of 36 patients who underwent arthroscopic patellar debridement for patellofemoral pain. All patients had isolated chondromalacia patellae noted during arthroscopic examination. No patient had a history of patellar instability or physical or radiographic signs of patellar malalignment. The chondromalacia patellae was classified as traumatic or atraumatic in origin. All patients had failed results after a minimum of 4 months of physical therapy before surgery, and all patients had grade 2 or worse chondromalacia patellae at the time of debridement. At the time of followup, patients were evaluated by questionnaire, Fulkerson-Shea Patellofemoral Joint Evaluation score, independent physical examination, and radiographs. Patients were also asked to subjectively score their knees preoperatively, at the time of maximal improvement postoperatively, and at the time of followup for comparison. Preoperative examinations and radiographs were compared with examinations at the time of followup. The most significant finding was the improvement in the overall joint evaluation score. The score for the entire group improved from a mean of 51.9 preoperatively to 78.8 at the time of maximal improvement and 75.3 at the time of followup. All but four patients subjectively thought that surgery had a beneficial effect. There were no observed changes in the preoperative and postoperative radiographs. Patients with traumatic chondromalacia patellae had 57.9% good or excellent results with surgery, and the patients with atraumatic cases had 41.1% good or excellent results with surgery, indicating that many patients who were improved by the surgery still had functional limitations.