Sentiment Analysis has played a crucial role in determining the effectiveness of the response and recovery efforts. In this paper, we present a survey exploring the application of sentiment analysis techniques in disaster management, focusing on their role in early warnings and disaster prediction using data from social media and crowdsourcing platforms. Extracting and categorizing emotional cues from online content during crises could help to interpret public sentiment to enhance early warning systems and disaster response strategies by harnessing machine learning techniques, such as supervised classification and deep learning models. The survey reviews various methodologies and tools employed in recent studies, focusing on their contributions to real-time disaster monitoring, sentiment-based classification of disaster events, and proactive management of crises. Insights drawn from these studies underscore the potential of sentiment analysis in leveraging online data to inform decision-making and improve resilience against natural disasters and other emergent events.