Abstract: The high prices of fruit and the hard work of the pickers create a need for automation in the fruit picking process. In this paper, a solution for the harvesting of the BabyGold9 peach variety is proposed. For this purpose, the main characteristics on which a correct automation depends are analysed and different solutions for the three main systems are discussed. According to the identification system, the YOLOv8s-P software is proposed, which will capture images using Real Intel Sense D43Di cameras. The navigation system is divided into two sub-systems, one in charge of mapping the environment and knowing the location and the other one who plans the routes to the peach. To tackle both tasks, the Gmapping software has been proposed with the T265 camera that will map the environment and the RRT* algorithm which is in charge of creating routes avoiding obstacles. A cost estimation and a performance evaluation will be presented in order to know what minimum requirements the robot should have in order to be able to replace human beings. Finally, a commercial end-effector and manipulator have been chosen for the picking system. These are a three-finger soft end-effector developed by the company SoftGripping and the UR5e robotic arm developed by Universal Studios. Finally, an explanation of how it should be tested and the necessary parameters to be analysed is given.