The aim of the present study was to understand what factors influence infants’ problem-solving behaviours on the multiple-string task. The main question focused on why infants usually solve the single string-pulling task at 12 months at the latest, whereas most 16-month-old infants still cannot solve the task when several strings are presented, only one of which is attached to the desired object. We investigated whether this difficulty is related to infants’ ability to inhibit their spontaneous immediate actions by comparing active and purely visual performance in this task. During the first part of the experiment, we assessed the ability of infants aged 16–20 months to solve the multiple-string task. The infants were then divided into three groups based on performance (a “failure” group, an “intermediate” group, and a “success” group). The results of this action task suggest that there were differences in infants’ performance according to their level of inhibitory control of their preferred hand. In the second part of the experiment, the three groups’ predictive looking strategies were compared when seeing an adult performing the task. We found that only infants who successfully performed the action task also visually anticipated which string the adult had to pull in the visual task. Our results suggests that inhibitory control was not the only factor influencing infants’ performance on the task. Furthermore, the data support the direct matching hypothesis (Rizzolatti and Fadiga, 2005), according to which infants need to be able to perform actions themselves before being able to anticipate similar actions performed by others.