IMPORTANCE: Repeated low-level red light (RLRL) therapy has emerged as a potential intervention for controlling myopia progression in children. However, its long-term effects on retinal photoreceptors remain relatively unknown. OBJECTIVE: To evaluate changes associated with RLRL therapy on cone photoreceptor density in children with myopia using high-resolution adaptive optics scanning laser ophthalmoscopy (AOSLO). DESIGN, SETTING, AND PARTICIPANTS: This retrospective multicenter cohort study analyzed data collected from January to March 2024, focusing on Chinese children with myopia. All participants were recruited through questionnaires. Cone density measurements were obtained from AOSLO retinal images. Children with myopia aged 5 to 14 years recruited from the pediatric ophthalmology clinic during routine eye examinations were included in the study and assigned to the RLRL group or the control group. Inclusion criteria were spherical equivalent refraction below -6.00 diopters (D) and best-corrected visual acuity ≥20/20. EXPOSURES: Cone density measurement with AOSLO retinal images. MAIN OUTCOMES AND MEASURES: Cone photoreceptor density was measured along 4 retinal meridians from central fovea to 4° eccentricity on AOSLO. Fundus abnormalities were assessed using AOSLO, optical coherence tomography (OCT), and fundus photography. Image evaluators were masked to group allocation. RESULTS: A total of 99 children with myopia were included in this analysis: 52 (97 eyes; mean [SD] age, 10.3 [1.9] years; 27 female [51.9%]) in the RLRL group and 47 (74 eyes; mean [SD] age, 9.8 [2.1] years; 25 male [53.2%]) in the control group. RLRL users showed decreased cone density within 0.5-mm eccentricity from the foveal center, most notably in the temporal region. At 0.3-mm temporal eccentricity, the RLRL group showed a mean difference of -2.1 × 103 cells/mm2 compared with controls (95% CI, -3.68 to -0.59 × 103 cells/mm2; P = .003). A total of 11 eyes exhibited abnormal low-frequency, high-brightness signals near the fovea. The odds ratio of abnormal signals in RLRL users compared with nonusers was 7.23 (95% CI, 1.15-303.45; Fisher exact test, P = .02). One participant had relatively small cystoid abnormalities on OCT in the ganglion cell layer, which resolved 3 months after discontinuing RLRL therapy. CONCLUSIONS AND RELEVANCE: Results of this cohort study suggest that RLRL therapy for at least 1 year was associated with reduced cone density in the paracentral fovea and other subtle retinal abnormalities in some children receiving this therapy for myopia control. These findings support the need for further research to evaluate longer-term safety of RLRL therapy in similar individuals.