The relentless growth of complexity and heterogeneity of CPS and IoT systems imposes major challenges on and offers new opportunities for design automation methods and tools. Intrinsic heterogeneity—the large span in size from small CPS appliances to societal-scale systems—enabled by IoT devices and the complex interaction among computational, physical, and human components makes the development of design automation tools necessary but very difficult. One hard challenge in automating heterogeneous CPS design flows is that they crosscut several discipline-oriented design verticals. Their integration requires new co-design methods, horizontal integration platforms for domain-specific models and tools, and the incorporation of data analytics tools for advancing the fusion of model- and data-driven design methodologies. The upcoming wave of autonomous systems incorporating learning-enabled components drives the need for a tighter integration of design-time methods and tools into operation. In the new generation of evolving, self-adaptive CPS, critical system properties must be ensured at design time as well as at run time. Formal modeling, synthesis, and verification tools are then needed that can provide both design-time and operation-time evidence for calculating assurance levels. Besides the fundamentally new design automation architectures, CPS and IoT application domains require unique modeling, analysis, simulation and synthesis tool components and efficient methods for rapid, inexpensive and semantically precise configuration of focused design tool chains. Advancement in this area needs comparative analysis of variability among design automation tool suites in a wide range of application domains.