In recent decades, there has been a notable surge in eco-friendly investments among prospective and current market participants. It has now evolved into an imperative standard for contemporary business decision-making. The critical challenge lies in the adept selection of investment portfolios that can effectively address both sustainability and profitability concomitantly. This task inherently involves a multi-dimensional nature, necessitating the consideration of various, sometimes conflicting, metrics. Consequently, this study proposes an advanced decision-making framework that amalgamates Data Envelopment Analysis (DEA) with Artificial Intelligence (AI). In contrast to prior research, which predominantly focused on determining investment portfolios based solely on target profitability, this study extends its purview to encompass sustainability considerations. This expansion aligns with the prevailing business trend that prioritizes fostering an eco-friendly atmosphere. While the DEA-based approach has demonstrated its supremacy in performance evaluation, it exhibits a notable limitation – a dearth of forecasting capability. To address this limitation, we introduce the Extreme Learning Machine (ELM), an AI methodology renowned for its superior predictive capacity. This integration is inspired by ensemble learning principles, aimed at harnessing complementary insights from diverse contributing models. Through the incorporation of this model, the role of market participants undergoes a profound transformation, shifting from retrospective and current monitoring to proactive future planning.