Background Screening for amblyopia at earliest is important for early treatment and better prognosis. This study aimed to evaluate the validity of uncorrected distant and near visual acuity (VA) and stereoacuity for screening amblyopia in grade-1 students in primary school in central China. Methods By stratified cluster sampling, 3112 grade-1 students from 11 Anyang primary schools were selected for the study. All the participants underwent uncorrected distant and near VA, stereopsis test, cycloplegic refraction, best corrected VA (BCVA), cover test, and ocular movement examination. VA was measured with a logarithm of the minimum angle of resolution (logMAR) chart. Stereoacuity was measured with the Lang II stereo card and TNO test. Amblyopia was defined as the BCVA less than or equal to 0.1 logMAR units of any eye in the absence of significant pathological abnormalities. The sensitivity, specificity, and positive and negative predictive value of uncorrected VA and stereoacuity for amblyopia were analyzed. Results Out of the 3112 eligible students, 2893 (92.96%) completed the examinations. The average age of the students was (7.10 +/- 0.41) years. Screened by distant VA with low cutoff (logMAR 0.1), high cutoff (logMAR 0.0), and near VA (logMAR 0.0), 31.64%, 73.18%, and 50.23% students were abnormal, respectively. Screened by stereopsis test, only 4.69% students were abnormal. Diagnosed by a senior pediatric ophthalmologist, 61 students had amblyopia. The sensitivities of distant VA with low/high cutoff and near VA were 92.31%, 100%, and 80.77%, respectively, whereas that of stereoacuity by TNO test was 15.38%. Simultaneous testing of either two of the three tests improved the sensitivity. Conclusions Distant VA test of high cutoff alone displays a high sensitivity but a low specificity. Simultaneous testing of distant VA of low cutoff and stereoacuity is a better choice to balance between sensitivity and specificity.