This study aims to explore customer sentiment and demand prediction models based on cognitive computing to improve companies’ understanding of customer behavior and optimize market strategies. By combining deep learning and natural language processing technology, this paper builds a sentiment analysis model that can effectively extract emotional tendencies from customer feedback, and a demand prediction model that combines sentiment analysis results and historical transaction data to predict future customer needs. Empirical analysis shows that the sentiment analysis model exhibits high accuracy, while the demand forecast model shows good predictive performance, proving the effectiveness of the comprehensive model in customer demand forecasting. In addition, this study also discusses the application potential of the model, providing companies with a new tool to better understand customer behavior and develop precise market strategies. Although there are certain limitations, such as the impact of the size and diversity of the data set on the generalization ability of the model, and the need to improve the interpretability of the model, this study provides a basis for future research in this field and points out possible research directions, including expanding data sources, exploring new data analysis techniques and algorithms, and enhancing the versatility and adaptability of models.