A study of 20,750 students in southern China aged 5–18 years (2018–2021) compared two criteria for detecting myopia in school screenings based on noncycloplegic measurements in at least one eye: SER ≤ −0.50 D only and SER ≤ −0.50 D plus age-impaired uncorrected visual acuity (UCVA >0.20 logMAR at age 5; >0.00 logMAR aged ≥6 years). The SER + UCVA criterion showed lower myopia prevalence estimates than SER alone, e.g. 48.89% versus 59.48% in 2021, which corresponds to a 15-21% relative overvaluation. The overestimation was greatest in children aged 5–6 years (relative difference >50%), where the prevalence in SER almost doubled, and decreased to 10–15% in adolescents (14–18 years). The prevalence of myopia by SER + UCVA significantly accelerated in children aged 7–10 years between 2019 and 2020. From age 10, prevalence by SER + UCVA was significantly higher in girls than in boys (all P < 0.001). The combination of SER and UCVA reduces the overestimation of myopia, especially in younger children.