Integrating Large Language Models (LLMs) into financial sentiment analysis offers a promising avenue for understanding market dynamics. This research proposes a novel approach that combines the outputs of multiple LLMs to assess the sentiment of financial social media posts, simulating a fuzzy system to handle the inherent uncertainty and subjectivity in financial discourse. The methodology involves selecting a diverse set of LLMs fine-tuned for financial contexts. Each model analyzes the same financial social media post, providing sentiment classifications. These outputs are then aggregated using fuzzy logic principles, assigning degrees of confidence to each model’s prediction to compute a final sentiment score. To validate the effectiveness of this approach, our study employs a dataset of financial social media posts, comparing the ensemble model’s performance against individual LLMs and traditional sentiment analysis methods. This research presents a method for financial sentiment analysis, integrating multiple LLMs and fuzzy logic, which would be useful for investors, financial analysts, and policymakers, as it allows for a more nuanced and accurate sentiment assessment.