Mechanoluminescent (ML) materials, which refers to a class of material that emits light upon being subjected to external mechanical stimuli, have been drawing attention due to their unique multifunctionality and potential applications in next-generation structural health monitoring. However, the practical use of ML materials has been limited by the lack of a universally acceptable framework for producing high-intensity ML composites and the challenges in fabricating complex 3D shapes. As a breakthrough, this paper presents a novel approach for producing SrAl2O4:Eu2+, Dy3+ particle-based ML composites using Vat photopolymerization (VP)–based 3D printing, and its process optimization via machine learning-based algorithm. In this study, multi-objective Bayesian optimization (MBO) with Gaussian process regression (GPR) as surrogate model is adopted to fine-tune three critical process parameters in VP process: ML particle content, layer thickness, and cure ratio, aiming for both strong ML properties and reduced printing time. GPR models the complex process input-output relationship, utilizing data collected from experiments. The Pareto-optimal solutions identified by MBO not only enabled the rapid production of high-performance ML specimens but also provided empirical insights into how process parameters affect the end product’s ML property and overall printing time. Furthermore, a micromechanical method for analyzing ML particle volume fraction dependence of the ML intensity is presented to analyze the result. Finally, the practical applicability of the proposed framework was validated by testing ML-based stress sensors and mechanical components using the optimized VP process.