In this paper, design, modeling, and control of a grip-based climbing robot are performed, which consists of a triangular chassis and three actuating legs. This robot can climb through any trusses, pipeline, and scaffolds structures and can perform any inspectional and operational tasks in the high height which decreases the falling danger of operation and increases the safety of the workers. The proposed robot can be substituted for the workers to decrease the risk of death danger and increase the safety of the operation. Since these kinds of infrastructures are truss shaped, the traditional wheel-based climbing robots are not able to travel through these structures. Therefore, in this paper, a grip-based climbing robot is designed to accomplish the climbing process through the trusses and infrastructures in order to perform inspecting and manipulating tasks. Hence, a proper mechanism for the mentioned robot is designed and its related kinematic and kinetic models are developed. Robot modeling is investigated for two different modes including climbing and manipulating phases. Considering the redundancy of the proposed robot and the parallel mechanism employed in it, the active joints are selected in a proper way and its path planning is performed to accomplish the required missions. Concerning the climbing mode, the required computed torque method (CTM) is calculated by the inverse dynamics of the robot. However, for the manipulation mode, after path planning, two controlling strategies are employed, including feedback linearization (FBL) and adaptive force control, and their results are compared as well. It is shown that the latter case is preferable since the external forces implemented on the end effector tool is not exactly predetermined and thus, the controller should adapt the robot with the exerted force pattern of the manipulator. The modeling correctness is investigated by performing some analytic and comparative simulation scenarios in the MATLAB and comparing the results with the MSC-ADAMS ones, for both climbing and manipulating phases. The efficiency of the designed controller is also proved by implementing an unknown force pattern on the manipulator to check its efficiency toward estimating the mentioned implemented forces and compensating the errors. It is shown that the designed robot can successfully climb through a truss and perform its operating task by the aid of the employed adaptive controller in an accurate way.