The technological advancements in the field of robotics over the last decade have significantly improved the quality of human life in terms of achieving critical tasks with the development of humanoid robots. In an attempt to assist humans in achieving complex tasks either in known or unknown environment humanoid robot plays an important role with reduced risk inclusion. Humanoid robots are designed to emulate the aspects of human behavior in order to accomplish some tasks. Being reaching close to the target object or performing required manipulation or grasping one of them. These tasks are considered to be very simple for humans, yet the same is not true for a robot. It requires consideration of many aspects such as knowledge of the environment, modeling of the robotic system for controlling the robot actions. Recent studies demonstrate that for humans most of this information is obtained by vision. Vision is one of the most powerful and important sensor for interacting with the environment as it provides detailed information about the environment which may not be available using combinations of other types of sensors. Humanoid robots are playing increasingly important roles in realworld tasks, especially when it comes to space applications, where explorative missions are generally very expensive and due to environment asperity, such mission would often be more suited to unmanned one with autonomously operating humanoid robot rather than manned one. Thus various visual servoing technique which involves the use of visual sensory feedback to control the motion of the robotic system is developed in recent years.Driven by this motivation, an attempt has been made in this work to develop a framework for implementing visual servoing on real hardware for the eye in hand camera-robot configuration and in simulation for the eye to hand half humanoid robot.Firstly this work presents the widely used existing visual servoing methodology.Secondly, this work present the visual servoing methodology for tumbling object. Thirdly,the methodology for inclusion of secondary task through task sequencing for addressing the joint limit constraint arising due to robot kinematics structure is presented. Finally the implementation of a combination of these techniques for approaching the target object as the main task and avoiding the joint limits as secondary task for the half humanoid robot is presented. ROS is used for the implementation of visual servoing modules.