There have been few studies done on the abnormal function of velopharyngeal muscles in unrepaired cleft palate infants. To examine and assess velopharyngeal movement before primary palatoplasty offers supposedly any valuable information for the successful operation and the restoration of excellent velopharyngeal function. We designed to investigate and analyze velopharyngeal movement before and after primary palatoplasty in 26 cleft palate infants with a fine nasopharyngeal fiberscope. We found three different patterns of velopharyngeal movement in unrepaired cleft palate infants when crying or strangulation reflex occurred: (1) posterior movement type (10 cases, 38.5 percent), where the soft palates moved only posteriorly and cephalically and did not move medially; (2) medial movement type (10 cases, 38.5 percent), where the soft palates moved only medially and did not move posteriorly or cephalically; and (3) posteromedial movement type (6 cases, 23.0 percent), where the soft palates moved both posteriorly and cephalically as well as medially. Postoperative velopharyngeal closure was classified into three patterns: (1) the soft palate type, in which the soft palate mainly operates; (2) the lateral wall type, in which compensational medial movement of the lateral pharyngeal wall is mainly observed; and (3) the mixed type, in which both the soft palate and the lateral pharyngeal wall operate. Also, we demonstrated a close relationship between velopharyngeal movement before and after primary palatoplasty in cleft palate infants. In total, 10 of 16 cleft palate infants with the posterior movement type or posteromedial movement type, in which posterior movement of the soft palates was observed before primary palatoplasty, postoperatively showed the soft palate type of velopharyngeal closure. On the other hand, only 2 of 10 cleft palate infants with the medial movement type, in which the soft palates did not move posteriorly but medially before primary palatoplasty, postoperatively showed the soft palate type of velopharyngeal closure. The Fisher's exact probability test clarified that cleft palate infants with the posterior movement type or posteromedial movement type were more likely to show postoperatively the soft palate type of the velopharyngeal closure compared with those with the medial movement type (p = 0.051). This is the first trial to examine velopharyngeal movement in unrepaired cleft palate infants. Our findings indicate the probability that velopharyngeal closure mechanism in repaired cleft palate infants is able to be predicted by velopharyngeal movement behavior before primary palatoplasty. Next, we must clarify a correlation between preoperative velopharyngeal movement and postoperative velopharyngeal function and speech outcome.