Hurricanes create severe disruptions to infrastructure, supply chains, and emergency response systems, particularly in coastal regions. Accurate forecasting of hurricane paths is essential for preparedness and resource allocation, but relying on a single predicted path can create high uncertainty because hurricanes may shift direction before landfall. This study presents a data-driven framework for generating multiple possible hurricane trajectory scenarios using historical hurricane track data covering 173 years, from 1851 to 2023. Historical hurricane paths recorded at six-hour intervals are clustered using K-means based on trajectory similarity. For each cluster, an average path is generated by averaging latitude and longitude coordinates, creating representative hurricane movement scenarios. As a storm progresses, its observed position is compared with historical tracks within a selected spatial radius. At each six-hour step, the radius is gradually reduced, and scenario probabilities are recalculated using the proportion of hurricanes assigned to each cluster within the current spatial window. This sequential process allows uncertainty to decrease as more storm-position information becomes available. Instead of directing all emergency resources toward one deterministic path, the proposed framework supports planning across multiple plausible paths with associated probabilities. The results show that early-stage hurricane movement contains several feasible paths, while later stages become more focused as probability concentrates around fewer scenarios. This approach supports pre-hurricane preparedness, resource allocation, and disaster response planning under uncertainty.