Any system that integrates digital control in a physical environment is an instance of Cyber-Physical Systems (CPS). The design, analysis, and implementation of reliable CPS naturally require the aggregation of all existing insights on discrete and continuous systems, yet our success has been limited by a lack of unifying methodology. I claim that logic and automated reasoning can provide a framework that is theoretically adequate and practically powerful to address the core problems in this emerging field. The framework should drive new progress in model-based design, automated synthesis, formal verification, and ultimately the complexity and reliability of the next generations of CPS. Along this line I present some of our results and projects, such as delta-decidability over the reals, the dReal solver, and automated analysis of realistic hybrid systems from automotive, aerospace, and medical applications.