PURPOSE: To evaluate the best-corrected visual acuity and stereoacuity gains in children >7 years of age and adults with unilateral amblyopia treated with a prototype virtual reality-based binocular amblyopia therapy. METHODS: In this randomized, double masked, cross-in clinical trial, patients at Boston Children's Hospital with unilateral anisometropic and/or strabismic amblyopia and history of prior amblyopia treatment failure were randomized to either a full-treatment group (8 weeks of binocular treatment using therapeutic software application in virtual reality headset) or a sham-crossover group (4 weeks of sham treatment followed by 4 weeks of binocular treatment). Amblyopic eye visual acuity and stereoacuity were evaluated at 4, 8, and 16 weeks' follow-up. RESULTS: The study cohort included 20 participants (10 females), with a median age of 9 years (range, 7-38 years). In the full-treatment group (11 patients), the mean amblyopic eye logMAR visual acuity at 16 weeks was 0.49 ± 0.26, compared with 0.47 ± 0.20 at baseline. In the sham-crossover group, it was 0.51 ± 0.18 at 16 weeks, compared with 0.53 ± 0.21 at baseline. Stereoacuity (log arcsec) was significantly improved, from 7.3 ± 2 at baseline to 6.6 ± 2.3 at 8 weeks (P < 0.001) and 6.7 ± 2.6 at 16 weeks (P < 0.001). No significant adverse events (diplopia, asthenopia, or worsening strabismus) were noted in either group. CONCLUSIONS: Although the virtual reality-based prototype for binocular amblyopia therapy did not significantly improve visual acuity in the amblyopic eyes of older children and adults, stereoacuity did significantly improve compared with baseline; improvements were clinically minute. However, larger studies are required to confirm the results.