Visual motor integration (VMI) is considered a key predictor of adaptive functioning in children with autism spectrum disorder (ASD). However, information regarding the VMI characteristics of children with ASD is scarce because existing assessment methods have focused on outcome measures while neglecting process measures due to the high failure rates in children with ASD. Therefore, the current pilot study was conducted to explore the appropriate indicators for the construction of a comprehensive VMI evaluation. First, five computer games were designed to provide enjoyable and appropriate tasks for use in VMI assessment while ensuring a high completion rate among the participants. Then, 30 children with ASD and 30 age-matched typically developing (TD) children between 4 and 6 years old were recruited to play the games. In this process, several quantitative indicators were applied to infer the children's visual perception, eye-hand coordination, and fine motor skills. Finally, the validity of the proposed method was verified using three classic machine-learning algorithms. The results revealed that the proposed method is effective in providing a comprehensive description of VMI deficits and can thus ultimately provide an objective basis for targeted interventions.