The popularity of peer-to-peer (P2P) housing services has increased the demand for understanding guests' emotional experiences and sentiments towards these services. This study aims to fill this gap by analyzing guests' emotional tendencies expressed in online reviews of P2P housing services. Our approach proposes a novel multimodal fuzzy model that addresses the limitations of previous studies, such as the inability to capture the complexity and subjectivity of emotions accurately. The model leverages advanced text processing technologies, including the valance-aware dictionary and sentiment reasoner (VADER), count-vectorization, and word embedding, to provide a more nuanced and objective analysis of guests' sentiment and emotional valence. Additionally, this study examines the impact of the COVID-19 pandemic on guests' emotional tendencies and P2P innovative strategies. The results of this study have the potential to enhance effective strategies for the P2P industry and advance the field of sentiment analysis in the P2P domain.