A key issue in the practical application of fluorescence imaging is the presence of a background signal detected during data acquisition when no target fluorescent material is present. Regardless of the technology employed, background signals cannot be completely eliminated, which limits the detection sensitivity of fluorescence imaging systems, especially for in vivo applications. We present a methodology to characterize the sensitivity of fluorescence imaging devices by taking the background effect into account through the fluorescent signal-to-background ratio (SBR). In an initial application of the methodology, tissuelike liquid phantoms with Cy5.5 fluorescent inclusions were investigated experimentally over a wide range of varying parameters, such as tissue absorption coefficient, scattering coefficient, fluorophore concentration, and inclusion location. By defining detectable and quantifiable SBR thresholds, empirical relations are established, and the sensitivity performance of Advanced Research Technologies's eXplore Optix using Cy5.5 is characterized.