This study deals with the problems to acquire the features of land and sea surface environments from multiple satellite sensor data. The Seto Inland Sea area of Japan characterized with remarkable population increase was selected as an intensive study area. An emphasis was put on environmental parameters such as land cover and surface temperature information form multiple sensor data. A modified maximum likelihood method using the multi-layer classification image was applied to LANDSAT/TM data of two different dates for producing a precise land cover change map of a regional scale. The capability for land cover change detection of NOAA/AVHRR data was investigated using the combined data set of AVHRR and LANDSAT/TM. The discriminative capability of the spectra of AVHRR was evaluated on the basis of the regression analysis between the land cover classes retrieved from TM data and the spectral feature of AVHRR. Fairly good correlation was obtained for the vegetation and non-vegetation class, which suggests a possibility for the detection of the land cover change due to urban development using NOAA/AVHRR data. While sea surface temperature was extracted from MOS-1/VTIR data by applying the split window method. These multiple sensor data were made into environmental data base combined with GIS data which can be effectively utilized for monitoring and assessment of natural environments.