One of the significant problem in robot kinematics is to optimizing the solution of inverse kinematics which deals with obtaining the joint variables in terms of the end-effector position and orientation and is difficult than the forward kinematics problem. As the degree of freedom of a robot increases the inverse kinematics calculation become more difficult and expensive. In industrial and manufacturing field the use of robot is an inevitable factor and thus the motion of its manipulator is necessary to express more efficiently and simply. Because of this reason kinematical approach is considered here. The Kinematics is the analytical study of geometry of motion of a robot arm and is of two types, Direct and Inverse Kinematics. Here the Neural Network approach is adopted and the algorithm employed is Recurrent NeuralNetwork. This paper proposes neural network architecture to optimize the inverse kinematics solution. In this paper the Recurrent Neural Network (RNN) algorithm is proposed to solve the inverse kinematics problem of a five degree of freedom robot end effector. This technique causes a decrease in the difficulty and calculations faced when using the traditional methods in robotics. Thus the optimized output is evaluated to ensure the efficiency of this approach. © 2017 IAEME Publication.