<p>Integrated computational materials engineering has enabled a paradigm shift in the approach to materials design. Using predictive process–structure–property models in an inductive manner can enable rapid design of composition, process, and microstructures, to achieve specific property objectives. However, in realistic complex engineering applications there tends to be competing objectives as well as complex process–structure–property relationships with multiple sources of uncertainty that can affect performance. The concept of robust design can be leveraged to reduce variability and achieve design objectives in the face of uncertainty. This work applies CALPHAD modeling to predict process–structure relationships and crystal plasticity finite element method to predict structure–property relationships for static and microstructure-sensitive multiaxial fatigue properties of Ti64. A framework for uncertainty–informed inductive design exploration is demonstrated, quantifying aleatory uncertainty resulting from microstructure stochasticity and epistemic uncertainty resulting from model form and model parameters, and propagating uncertainty through process–structure–property linkages, utilizing Gaussian process machine learning surrogate models to accelerate the propagation of uncertainty from microstructure to multiaxial fatigue properties. The results show the advantages of considering uncertainty, providing unique optimal and robust design outcomes.</p>