Purpose: to evaluate the accuracy of the "Assessment sucking scale to newborn in breastfeeding". Methods: this is a diagnostic study. The "Assessment sucking scale to newborn in breastfeeding" was drawn from an existing instrument, has been validated. The study included 152 mothers / newborn hospitalized in rooming in care. The study was divided into two stages, the first stage adaptation of the instrument and the second validation of the "Assessment sucking scale to newborn in breastfeeding" itself. For comparison between the instruments, it was considered as the gold standard "Protocol assessment of breastfeeding" proposed by UNICEF. Statistical analysis was applied to test the association chi square (first stage) and building the "Receiver Operating Characteristics" curve (second stage). Results: in the first phase, the "Assessment sucking scale to newborn in breastfeeding" was drawn from the behavior of newborns who were associated with breastfeeding, consisting of items posture, movement and cupping tongue and jaw movement. However, in the second stage, it was found that "Assessment sucking scale to newborn in breastfeeding" has a poor accuracy and is not a predictor of the behavior of the newborn in the breastfeeding. Conclusion: the "Assessment sucking scale to newborn in breastfeeding" cannot be considered a predictor of behavior newborn in the breastfeeding.