Many good experimental results in noninvasive measurement of glucose have been achieved. However

lack of theoretical support counteracts their extensive acceptance. After analyzing the two kinds of effect on optical properties and further influence on light propagation induced by the variation of glucose concentration

the relationship between detection limitation of glucose concentration and source-detector separation and signal-to-noise ratio of measuring system was derived. Referring to skin’s layered model

signal-to-noise ratio was modified accordingly. Monte Carlo simulations performed in a skin-layered model show that the percent of effective photons from dermis layer can be larger than 50% by selecting optimal source-detector separation

and that the requirement of FDA for noninvasive glucose sensing could be satisfied if the instrument SNR of 10000 is supposed.