INTRODUCTION: cerebral palsy (CP) has the characteristic of causing changes in posture and movement that hamper the achievement of functional activities. In the face of motor disabilities, rehabilitation becomes essential and is an option based on motor learning. However, it is important to research the motor learning process in individuals with CP to make the organization of treatment programs more effective. OBJECTIVE: Analyse the motor learning in children with CP. METHOD: For the realization of this work, an experimental group (EG) and a control group (CG) was used, comprised of four children matched in relation to gender (one female and three masculine) and age (between seven and twelve years). The task was to conduct a path into a maze in the shortest time possible. The work consisted of two phases, being initially an acquisition phase (AQ) and then transfers (immediate-IM; short-term-ST and long-term-LT). RESULT: It was found that there was no statistical significance difference between AQ and transfers evaluated with the following values: IM (z= -1.83, p=0.07), ST (z= -1.83, p=0.07) and LT [EG (z= -1.83, p=0.07) and CG (z= -1.46, p=0.14)]. CONCLUSION: In the process of maze task learning, when analyzing the results between phases AQ and transfer, significant difference was not observed, which means that individuals with CP showed learning capacity through task adaptation equivalent to individuals without cerevral palsy.