Birth to the age of 3-5 y.o. is the critical period of visual acuity development in children. Preferential looking is an early eye examination to detect visual impairment in children and has been known as an objective test for children below 5 y.o. The purpose of this study was to examine the children’s accuracy in answering the computer-based preferential looking test. This research was conducted using a cross-sectional method. The subjects used were children aged 3-5 y.o. who were taken by convenience sampling technique in four urban kindergartens in Yogyakarta, Indonesia, and had met the inclusion and exclusion criteria. The visual acuity measurement technique used was a preferential looking with hemispatial Gabor patch written in Python application. The eye tracker was used as a control. The analytical method used was the one-way ANOVA method to see the significance of differences in overall eye accuracy. Mean Lea’s Chart visual acuity (LogMAR) of the eye tracked group: 0.21 ± 0.14 and the no eye tracked group: 0.28 ± 0.13. The accuracy of the eye tracked group (n=38 eyes) for the 3, 4, and 5 y.o. groups were: 81.14 ± 14.75, 75.98 ± 13.13, 81.00 ± 10.71, respectively (p= 0.459). Comparison between the eye tracked group and the no eye tracked group, the accuracy (n=38 eyes) of eye tracked group: 78.90 ± 12.11 and no eye tracked group (n=38 eyes): 86.21 ± 3,83 (p= 0.001). The children’s accuracy when performing the preferential looking test was remarkable considering their early age. Therefore, this method is applicable to a basic examination to get objective visual acuity in preschool children.