The rubidium content of infant formulas was determined by flame photometry using the atomic absorption spectrophotometer (AAS) with air-acetylene flame. In this work, the optimum instrumental conditions for rubidium analysis were determined: the burner height of AAS, the acetylene pressure, the hydrochloric acid concentration as a measuring solution and so on. The effects of coexisting ions expected from matrix components of infant formulas were tested; the results showed no interferences of Na⁺, Ca²⁺, Mg²⁺, PO₄³⁻ (∼250 μg/ml) and K⁺ (∼25 μg/ml). In addition, a variation of the sample amount (0.1∼1 g) for the analysis did not interfere with the determination of the rubidium content of the sample. Rubidium was determined as follows: A sample (0.3 g) in a beaker (50 ml volume) was decomposed by heating with a nitric-perchloric acid mixture on a hot-plate; the solution was then diluted to 100 ml with water after the addition of 2 ml of 6 mol/l hydrochloric acid. Standard rubidium was recovered to be 100 ± 0.9 (n = 20) from an infant formula sample. This method was applied to 24 infant formula samples, resulting in the fact that there was a significant difference in the rubidium content between 10.7 ± 2.9 μg/g (n = 14) in follow-up formulas and 5.35 ± 1.71 μg/g (n = 10) in young infant formulas (p<0.05).