Human-Machine Cooperation has been studied for more than thirty years in many domains such as transport, crisis management or manufacturing and hospital systems. Human-Machine Cooperation principles are now applied to complex industrial processes to address the increasing Human needs to keep their abilities to supervise and control such processes, when taking advantages of the increasing capacities of autonomous “intelligent” systems, including robots. The interest in using the models proposed by Human-Machine Cooperation is the ability to conduct a precise analysis of the functions that have to be performed by human operators and/or machines, in order to construct a planning based on those functions and their allocation to each of those actors. Planning can be simulated during task execution, and then adapted according to unexpected events, as they could appear with a process as little predictable as the disassembly process. The present communication aims at bringing some methodological bricks to design such a Human-Robot system managing constraints linked to the disassemble operations presented in this paper. Two experimental systems are then presented.