This article describes a model-based approach for defining and refining process parameters in dynamically changing, smart manufacturing environments. This approach uses equation-based models to predict how part quality will respond to changes in that environment. The results from these models provide the major inputs into a process-parameteroptimization technique, which is used to set the values for various process parameters. In developing these models, we integrated various concepts from process improvement frameworks, such as Define-Measure-Analyze-Improve-Control and Monitor-Analyze-Plan-Execute-Knowledge, with techniques from model-based engineering. After describing the approach, we demonstrate its use in an additive manufacturing process example.