To conduct a large retrospective study of screening refractive error in young children. This retrospective study included children aged from 4 months to 8 years in Daxing District, Beijing, who underwent refractive examinations without cycloplegia. It included a cross-sectional assessment of refractive error screening for all children, and a longitudinal component for a subgroup with data available for two to five visits. A total of 14,987 children were included in the cross-sectional study. In the group <1 year of age, the percentage of children with a spherical equivalent (SE) >+2.00 D or with cylinder <-1.50 D was 15.25% and 33.24%, respectively. These were significantly higher than for the 1- to 4-year-old group (SE 8.1% higher, cylinder 13.2% higher) (χ2 = 53.57, p < 0.001; χ2 = 790.39, p < 0.001). Furthermore, 34.83% of children in the 0-year-old group had amblyopia risk factors (ARFs). In the 4-year-old group, boys had a significantly longer axial length (AL) than girls (differences in the right and left eyes were 0.53 and 0.56 mm, respectively; z = 5.48 p < 0.001, z = 5.80, p < 0.001). AL increased with age, while the AL difference between boys and girls remained stable at 4-8 years of age. The percentage of children aged 5-8 years with myopia in 2020-2021 was significantly higher than that in 2018-2019 (H = 12.44, p = 0.006). In the longitudinal study of 4406 children (up to 12-month follow-up), annual changes in SE were -0.27, -0.06, 0.19 and 0.13 D between 0 and 3 years, and -0.38, -0.58, -0.70 and -0.75 D between 5 and 8 years. Children's refractive error varied significantly from ages 4 months to 1 year, with a high proportion having ARFs. Children aged 5-8 years showed a trend towards myopia. The prevalence of myopia in the cross-sectional analysis in 2020-2021 was greater than in 2018-2019. Screening refraction changed minimally over a 12-month period for children aged 1-3 years, but became more myopic for children aged 5-8 years.