The present study was designed to examine the effects of two kind of feedback presentation, namely KR and KP, on movement pattern and accuracy of a Badminton serve in children with autism. We used a causal-comparative method in the current study. Sixty children with autism with an age range of 7 to 12 years from special schools were selected based on a convenience sampling method and were randomly and equally assigned into four groups including KR, KP, KR+KP, and control groups. The motor task in the present study included the badminton serve, in which the movement patter and accuracy were measured as the dependent variable. The children participated in the pretest including 10 services, acquisition phase (5 training blocks, each of which included 10 services), and the retention test with 10 services. Respective feedback was provided before each practice block. We used ANOVA to analyze data. The results showed that both KR and KP feedback improve both the pattern and the accuracy of movement better than the condition without feedback. In addition, KP had better effects on the movement pattern and KR had better effects on the movement accuracy. Finally, children who were in the combination group performed better than all the groups in both execution of the movement pattern and service accuracy. Children with autism benefit from feedback to learn novel motor skills, indicating that they may have the necessary mechanisms to learn new skills through feedback.