Human integration into cyber-physical systems, such as the smart grid, smart transportation systems, and crowdsensing, while possibly contributing to the efficiency and sustainability of the system, introduces the possibility of externalities due to the actions of various participants who act strategically to further their own interest. In order to ensure desired system performance in spite of such strategic decision making, the system operator may need to design appropriate incentives for the users. This dissertation presents some techniques for incentive design in some particular problems modeled from practical cyber-physical systems and studies the behavior of the participants in response to the mechanism. In particular, we focus on the role of interest misalignment as well as information asymmetry among the participants in these systems. We focus on three settings: Demand response program in a smart (power) grid, distributed estimation in crowd sensing applications, and task allocation in a general multi-agent system. We design mechanisms to mitigate strategic behavior of the participants and address the challenges specific to each system, while satisfying certain properties such as individual rationality, incentive compatibility, and social optimality. We also study the gap between mechanism design and classical cost allocation for an incentive design in a general multi-agent system and propose future directions as well as several extensions of the presented work.