Polylactic acid, as a biodegradable material, has broad application prospects in packaging, medical devices and other fields. However, the insufficiency of its mechanical properties remains the main bottleneck restricting its large-scale promotion and application. To enhance the mechanical properties of polylactic acid and achieve intelligent control of the modification process, this paper proposes a performance prediction and process optimization method based on the random forest algorithm and constructs a data-driven intelligent configuration system for polylactic acid modification. In the study, 19 input features covering process parameters such as modifier types, addition amounts, relative molecular mass, and glass transition temperatures, along with two output targets, tensile strength and elongation at break, were selected to establish and train a random forest regression model. The results show that the optimized model demonstrates high prediction accuracy in both tensile strength and elongation at break performance indicators. The residual distribution is concentrated, the fitting trend is good, and it has strong nonlinear modeling ability and feature redundancy tolerance. It is stable in interpreting data variance and controlling prediction errors, and can meet the requirements of practical engineering applications. Further analysis of the importance of features revealed that the addition amount of modifiers, glass transition temperature and relative molecular mass are the key factors affecting the mechanical properties of polylactic acid, and they have a significant effect on the performance response under the combination of multiple variables. In the modeling process of this study, a sustainably updated dataset structure of polylactic acid modification was established. Subsequently, database expansion and model retraining can be achieved by supplementing new samples or features, thus demonstrating good scalability and sample adaptability at the data level. It provides a standardized, replicable and scalable data-driven solution for the performance prediction and process parameter optimization of bio-based materials.