In this study, we discuss the marine forecast zone by analyzing the characteristics of wind and wave produced by numerical model. In order to examine the spatial characteristics of marine weather, clusters were spatially classified over the study area through k-mean clustering method. Cluster analysis is conducted for sea surface wind speed and wave which is the most important variables in the marine forecast. Optimized cluster number among several clusters is chosen as 6 classes using mean silhoutte value. Cluster analysis results for wind speed (WS) cases is similar to mean distribution of sea surface wind, but wave height (WH) case is consistent with bathymetry. And cluster analysis which is considering wind and wave height simultaneously is analogous to wind speed cluster analysis. Mean wind speed and wave height of each cluster is increased away from the coastline. Coastal forecast zone which is classified in this study is more closer than current operational marine forecast zone. The cluster classification results are similar to each other in the Yellow Sea and Southern Sea of the Korean Peninsula.