Background Several Medial Patellofemoral Ligament (MPFL) reconstruction techniques have been developed, and those with soft tissue fixation are often preferred in children because they allow the growth cartilage to be preserved. Nevertheless, the recurrence rate of patellar dislocation varies widely from one series to another, with no clear superiority of one technique in the pediatric setting. The objectives of this study were to compare the results of two tendon graft fixation techniques (tendon-tendon fixation and anchor-screw fixation) by analyzing: 1) the rate of patellar dislocation recurrence, 2) clinical outcomes, 3) tourniquet time and 4) complication rate. Hypothesis The two tendon graft fixation techniques used in MPFL reconstruction are equivalent in terms of the patellar dislocation recurrence rate. Patients and methods This is a retrospective comparative study including 57 patients with a median age of 14 years (12–15 years) who underwent MPFL reconstruction between 2016 and 2020. The tendon graft was fixed upon itself, after passing through a patellar tunnel (Group A: tendon-tendon fixation; n = 29) or by two anchors and an interference screw (Group B: anchor-screw fixation; n = 28). The preoperative radiographic data were comparable in the two groups: patellar height [A: 1.3 (interquartile range (IQR): 1.2–1.4) / B: 1.2 (IQR: 1–1.4) (p = 0.21)], tibial tuberosity to trochlear groove (TTTG) distance [A: 16 (IQR: 13–19) / B: 13.5 (IQR: 11.5–18.8) (p = 0.12)], patellar tilt [A: 25 (IQR: 20–35) / B: 24.5 (IQR: 21–32) (p = 0.93)]. For each technique, the rate of patellar dislocation recurrence, clinical and functional results (Kujala score, Marx activity score, Lille patellofemoral score), complications (pain, stiffness, revision) were analyzed. In addition to MPFL repair, 13 patients (2 in Group A, 11 in Group B) underwent additional orthopedic procedures to enhance patellar stability. Results no patients were lost to follow-up and the median follow-up was 30 months (IQR: 20–38). The dislocation recurrence rate was higher in Group A, 6.9% (2/29) compared to none in Group B. The clinical results were comparable for the two groups with a Kujala score [A: 94 (IQR: 89–100) / B: 92 (IQR: 87.5–94.5) (p = 0.12)]; Marx score [A: 10 (IQR: 7–11) / B: 9.5 (IQR: 7.5–12) (p = 0.89)] and Lille patellofemoral score [A: 97 (IQR: 91–100) / B: 94 (IQR: 90–98) (p = 0.21)]. The tourniquet time was shorter in Group A than in Group B, 61 minutes (IQR: 52-71) versus 85 minutes (IQR: 55-115) (p = 0.024) excluding additional orthopedic procedures. The complication rate was 17.2% (5/29) in Group A (dislocation n = 2, stiffness n = 2, ATT (anterior tibial tuberosity) revision with screw removal n = 1) and 10.7% (3/28) in B (pain n = 1, ATT revision with screw removal n = 2) (p = 0.35). Conclusion Clinically, anchor-screw fixation appears to reduce the risk of patellar dislocation recurrence but this could not be statistically tested. On the other hand, the two techniques are comparable in terms of the functional results. Level of evidence III; retrospective case-control study.