Abstract : As capital flows increasingly follow responsible investing principles, the need for reliable ESG scores increases. ESG scores provided by rating agencies may be incomplete, inconsistent, and incomparable. This research aimed to present techniques to improve ESG scores for multidimensional decision-making. First, the Linguistic Ordered Weighted Geometric Aggregating LOWGA operator was used to obtain the Fuzzy ESG score of the firms where linguistic attributes were added to ESG scores generated by Refinitiv. This was followed by ranking firms based on the Fuzzy ESG score, stock return, and volatility by using the TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method. Adding linguistic attributes to the ESG scores using a LOWGA operator resulted in significant changes in the ESG scores. Considering that the linguistic attributes indicated how good the ESG scores were from an investor's perspective, the analysis demonstrated that using linguistic attributes can significantly alter ESG scores. The changes also make the generated ESG scores more beneficial to investors. However, it is the TOPSIS ranking that is most impactful on the utility of the ESG scores. TOPSIS ranking using proximity index to the positive ideal solution significantly altered the ranking of firms based on ESG scores. Since the ranking of firms is used in making decisions for responsible and sustainable investment, the change in ranking meant that the investment decision would be significantly improved when the TOPSIS ranking was used. The analysis was especially important because it incorporated information on risk and returns in the computation and analysis of ESG scores of firms. Risk and return are both essential factors of consideration in investment appraisal. The research recommends improvements to the process and methods of computing the ESG scores using the LOWGA operator and TOPSIS to include more information for decision-making, including risk and return measures.