Collaborative robotic manipulators-based precision farming robots need to accurately reach a target plant and safely handle poses performing plant-centered operations like weeding and irrigation with minimal plant-robot interaction. Unwanted contacts could inhibit the potential growth of plants. In this paper, we propose a human behavior-based hybrid inverse kinematics solver reinforced with a task planning algorithm inspired from the method of making ‘Flower Carpets for the commercial greenhouses to improve the reachability, safety (of plants) and success rate of collaborative manipulation-based tasks in greenhouse environments. A unified vision-based approach with feedforward neural network is used to approximate plant related geometric features like root location, area of spread and plant bed layout to implement an end- to-end manipulation-based precision robot is discussed with implementation results. Furthermore, the system was implemented and evaluated on gazebo simulation as well as a real mobile manipulator robot that was made to operate in a greenhouse setup. The proposed Hybrid IK planner is found to have a 32.27% increase in success rate in comparison to the standard IK solver. Unwanted plant interactions per task were reduced by an average of 0.45 Contact Factor. Similarly, the execution time was also reduced by an average of 58 s.