Seagrass and seaweed beds are very important for fish’s growth, but the beds are sensitive to environmental changes and human activities. Coastal zone monitoring is necessary for the conservation of seagrass and seaweed beds. Satellite data are used for the coastal zone monitoring. There are large atmospheric effects at the satellite data, and low altitude observations using UAV have the benefits of small atmospheric effects. Hyperspectral image data are powerful tools for determination of seagrass and seaweed species, and we want to develop a hyperspectral imaging sensor on UAV for the environmental study of coastal zone. In our previous studies, a hyperspectral array sensor system for an unmanned radio-controlled helicopter was developed. The specifications of the hyperspectral sensor were 400-1000 nm spectral range, 121 bands, 5 nm spectral resolution, and 11 kg weight. This system is too heavy for ordinary UAV and very expensive system. The hyperspectral imaging sensor on UAV has to be compact, lightweight and low cost. Last year, at the 34th ACRS, we presented the measurement system, and it could observe only nadir and the observed data were on flight line. Now we are making the prototype of hyperspectral image sensor, which can get area data using swing mirror. The spectral data of target ground or sea-surface are gathered at one side of glass fibers bundle. At the other side of glass fibers bundle, the spectral data are sequentially given to seven glass fibres by the swing mirror. The swing mirror is controled by a stepping motor and the swing speed of the mirror are variable. Eight Mini-Spectrometers manufactured by Hamamatsu Photonics are prepared. Each Mini-Spectrometer is connected one glass fibre. One Mini-Spectrometer is used for sky light monitoring and sevmeters are used for spectral data of target earth surface. The total weight of the the prototype of hyperspectral image sensor is about 5 kg. The system loaded to unmanned airplane in this summer and we got early phase results.