In addition to designing new donor (D) and/or acceptor (A) molecules, the optimization of experimental fabrication conditions for the organic solar cells (OSCs) is also a complex, multidimensional challenge, which hasn’t been theoretically explored. Herein, a new framework for simultaneous optimizing D/A molecule pairs and device specifications of OSCs is proposed, through a quantitative structure-property relationships (QSPR) model built by machine learning. Combining the device parameters with structural and electronic variables, the built QSPR model achieved unprecedentedly high accuracy and consistency. Additionally, a huge chemical space containing 1,942,785 D/A pairs is explored to find potential synergistic ones. Favorable expereimental parameters such as root-mean-square (RMS) and the D/A ratio (DAratio) are further screened by grid search methods. Overall, this study suggests the feasibility to optimize D/A molecule pairs and device specifications simultaneously by enabling better-informed and data-driven techniques and this could facilitate the acceleration of improving OSCs efficiencies.