SCIENTIFIC ABSTRACT The lack of the realistic mass-loss formula has been the main cause of the uncertainty in the theory of the stellar evolution. With known distances, metallicities and with a wide range of age, the star clusters in the Magellanic Clouds offer a unique opportunity to understand the role of mass- loss in the late stage of stellar evolution. We propose here to measure the infrared excess of individual AGB stars in these clusters. For the first time, the reliable mass-loss rate will be obtained along the AGB sequence of low to intermediate mass stars. A crucial test of the evolutionary calculation will be given by the comparison between the mass-loss rates derived from these observations and the mass-loss formula assumed in the theory. OBSERVATION SUMMARY CAM images with a PFOV of 3" will be taken for the rich stellar clusters in the Magellanic Clouds using three LW filters, LW1, LW2, and LW10. The expected signal to noise ratios are about 16 at 4.5 microns (LW1), 30 at 6.75 microns (LW2), and 30 at 11.5 microns (LW10), respectively for flux levels of 1 mJy, i.e., of bright AGB stars. We will use ISOCAM micro scan mode (3 x 3 or 4 x 4 raster with scan step being 48") to cover an individual cluster whose size is typicaly 1 to 3'. The scan direction is the spacecraft y axis. As an additional program, LINKED Observations with PHOT-P are proposed. Photometry at 25 microns through 23" aperture and at 60 microns through 52" aperture will be done for 5 brightest AGB stars in NGC 1783 and NGC 419, respectively, found in the course of CAM observations. Expected signal to noise ratios are about 9 at 25 microns for flux level of 50 mJy and 4 at 60 microns for flux level of 10 mJy with 128 sec. integration time. CONCATENATION: In order to obtain accurate color-color and color-magnitude diagram, CAM observations of each cluster with three filters should be concatenated because AGB stars often show variabilities. LINKED OBSERVATIONS: It is expected that very red, heavily obscured AGB stars will be found in the CAM observations. Since these stars may have great importance in understanding the mass-loss phenomenon itself and its effects on the late stages of stellar evolution, we propose to observe them photometrically with PHOT-P 25 and 60 micron bands. From CAM observations, we will select 5 infrared brightest stars in NGC 1783 and NGC 419, respectively, for this purpose. As the positions of these stars are not known in advance, the PHOT observations (OSN 4-8, 12-16) have to be linked to the CAM observations. Photometry should be carried out in several orbits after the CAM observations.