Enterobacter sakazakii, designated as an unique microbial species in 1980, may cause bacteremia, necrotizing enterocolitis and infant meningitis. The distribution and the thermostability of E. sakazakii in unprocessed ready-to-eat (RTE) agricultural products of 252 and in 25 powdered infant formulas (PIF) were analyzed. Eighty one, 50, 43, and 47% of brown rice, pumpkin, potato, and carrot samples, respectively, had aerobic plate counts (ARC) in the range of 5 log CFU/g or more. Almost all the other products sampled had APC of approximately 2 log CFU/g. Fifty three, 75, 67, and 68% of banana, pumpkin, soybean, and carrot had Enterobacteriaceae counts approximating 3 log CFU/g. Sixty six percent of the brown rice tested had Enterobacteriaceae counts approximating 5-6 log CFU/g. E. sakazakii was isolated from 3/25(12%), 4/23(17%), 1/24(4%), and 1/27(4%) of PIF, brown rice, laver, and tomato samples, respectively. D-values were 3.52-4.79 min at 60 and $D_{60}-values$ were similar as the isolates reported. Thermal inactivation of four thermovariant E. sakazakii strains during the rehydration of PIF with hot water were investigated. At $50^{circ}C$, the levels of E. sakazakii decreased one log CFU/g for 4-6 min and thereafter the levels remained stable for 20 min. At $60^{circ}C$, inactivation by about 2 log CFU/g occurred for 20 min. Therefore, the unprocessed agricultural products might be a source of contamination for PIF when used as an ingredient after drying and pulverization. Rehydration of PIF for infant feeding with a water temperature of $60^{circ}C$ rather than $50^{circ}C$, as recommended by the manufacturers, may be helpful in the reduction of potential E. sakazakii risk.