In this article, we present a learning model that can control the kinematics motion of a simulated anthropomorphic arm in reaching and grasping tasks of a static prototypic object placed behind an obstacle of varying position and size. The network, composed of two generic neural network modules, learns to combine multi-modal arm-related information (trajectory parameters) as well as obstacle-related information (obstacle size and location). We based our simulation on the Via Point notion, which postulates that the reach motion planning is divided into successive positions of the arm. In order to determine these particular positions, some specific parameters have been extracted from an experimental protocol and constitute the pertinent parameters to be integrated into the model. This net of neural net determines the total path able to reach and grasp the prototypic object while avoiding an obstacle.