Background <p>Retinal degenerative diseases (RDD) cause irreversible vision loss due to photoreceptor (PR) loss. Stem cell–derived PR precursors hold promise for retinal repair, but genetic labeling limits clinical translation. Surface marker–based sorting offers a safer alternative.</p> Methods <p>We analyzed single-cell RNA sequencing (scRNA-seq) datasets of human retinal organoids to identify PR-specific surface markers and validated them by immunofluorescence and flow cytometry. Cells with a CD9⁻CD81^mid (negative for CD9 and intermediate levels of CD81) phenotype were isolated via fluorescence-activated cell sorting (FACS) and transplanted into the subretinal space of RCS rats. Post-transplantation survival, integration, differentiation, and therapeutic efficacy were assessed through immunofluorescence, quantification of outer nuclear layer (ONL) thickness, scotopic flash electroretinography (fERG) and light/dark box test.</p> Results <p>Among the identified markers (STX3, PARD6B, PROM1, CD9⁻CD81^mid), the CD9⁻CD81^mid phenotype effectively enriched PR precursors with high purity and low proliferative activity. After transplantation, these cells survived, integrated into the host retina, and differentiated into cones and rods. Treated rats showed preserved ONL thickness and significantly improved visual responses for up to 12 weeks.</p> Conclusions <p>We established a clinically translatable, non-genetic strategy for enriching human PR precursors using the CD9⁻CD81^mid marker combination. The enriched cells demonstrate the capacity for photoreceptor differentiation and functional rescue in a preclinical model of retinal degeneration highlighting their potential for cell-based therapy in RDD.</p>