Due to the extremely weak blood glucose signal carried by photons transmitted through human tissues, the sensitivity of near-infrared spectroscopy for measuring blood glucose is low, making it difficult to achieve high accuracy. Therefore, ongoing research has attempted to enhance sensitivity by optimizing measurement wavelengths, distances, and other factors. However, these studies have often focused on variations in tissue absorption and scattering coefficients caused by blood glucose while neglecting the impact of the optical properties of the measured tissue itself on sensitivity. They lack comparisons between different anatomical locations to optimize measurement sites. The optical parameters of the tissue itself affect absorption and scattering changes through their influence on optical path length, affecting the coefficient of the interaction between scattering and intensity changes. Therefore, a more reasonable approach is to comprehensively consider both factors when determining glucose measurement sensitivity. This study selected four concentrations of intralipid solution (2%, 5%, 10%, and 20%) as tissue phantoms to simulate human tissue. Using Monte Carlo simulation, the study investigated glucose sensitivity within the 1 000~1 660 nm wavelength range, considering glucoses absorption, scattering, and combined effects in the four solutions. The study also explored the relationship between the sensitivity of each component and its optical parameters. The results indicate that the strongest glucose signal was detected in the 20% intralipid solution, which had the highest scattering coefficient. Based on this, the study provides a basis for site selection to achieve higher glucose sensitivity. Additionally, analyzing glucose signals within the 1 000~1 660 nm wavelength range, it was found that in the 1 000~1 350 nm range, the absorption effect of glucose could be generally ignored, and signal differences mainly stemmed from changes in scattering. In the 1 350~1 660 nm range, scattering and absorption jointly influenced the signal, with scattering contributing more significantly. The optimal measurement wavelength for scattering was around 1 450 nm while considering the combined effects of scattering and absorption, the optimal measurement wavelength was around 1 400 nm. Finally, to validate the theoretical analysis further, measurement experiments were conducted on the four solutions using six typical wavelengths within the 1 000~1 660 nm range. The results showed that the wavelength with the highest glucose signal in all four solutions was 1 409 nm, and the sensitivity of the glucose signal in the 20% intralipid solution was the highest. It indicates good agreement between experimental results and theoretical analysis. This study provides valuable insights for selecting appropriate measurement sites and wavelengths for non-invasive blood glucose detection in the human body.