Accessibility to cloud-free optical sensor images is essential for large-area monitoring of land and forest cover changes. In this study, the acquisition probabilities of cloud-free images were analyzed using MODIS cloud mask products from 2000 to 2008 in Southeast Asia. The daily cloud masks were summarized into monthly acquisition probabilities for cloud-free images over the period at a spatial resolution of 1km. The mean annual acquisition probability profiles were extracted averaging nine years' observation. Unsupervised clustering was conducted for zoning of the acquisition probabilities using the mean annual profiles. Annual variations in the acquisition probabilities were examined by the standard deviations calculated for each month and comparisons of the mean annual profiles of the whole period and individual years. The distributions of annual acquisition probabilities in forested areas were different in each country. These results suggested that selection of suitable methods and data allowing for the spatial and temporal differences in the acquisition probabilities is necessary for periodic large-area monitoring.