The current reliance of electric batteries on transition metals, which require extensive mining and contribute to greenhouse gas emissions, presents a significant environmental challenge. Redox-active organic batteries offer a promising alternative with a lower carbon footprint, but their development is often hindered by issues such as high dissolution rates and low electronic conductivity, requiring time-consuming and expensive experimentation. To address these challenges, we harness the power of organic battery informatics, utilizing advanced machine learning (ML) techniques to accelerate the discovery and optimization of suitable redox-active organic materials for battery applications. Our data-driven approach employs proxy properties and transfer learning to enhance model accuracy and efficiency, outperforming random searches and baseline models. By training our ML model on a comprehensive dataset of organic batteries and their associated properties, such as voltage and specific capacity, we can effectively screen a vast library of candidate polymers to identify those with the highest potential for battery applications. These top candidates are passed on to experimentalists for validation and further development, paving the way for a new generation of sustainable and high-performance organic batteries.