Circulating blood volume (CBV) is important for disease evaluation and operation monitoring of patients with cardiovascular and cerebrovascular diseases in clinical application as the main hemodynamic parameters. Indocyanine green (ICG) was used as a tracer in the pulse dye densitometry, and we detected CBV noninvasively in vivo by establishing the spectral curve of the diluted excreted ICG. In clinical application, the accuracy of CBV detected with pulse dye densitometry is lower than the expected value owing to the influences of blood oxygen fluctuation and background light. To solve the problem, this paper researched circulation blood volume detection by modified pulse dye densitometry. The specific process was as follows. We injected the ICG reagent into the vein in the elbow of patient, and collected the transmission spectrum signal and the background photoelectric signal respectively using the photoelectric sensor. Then we conducted differential operation on the collected data to eliminate the influences of blood oxygen fluctuation and background light and established an accurate ICG spectral curve. So CBV and other hemodynamic parameters could be calculated finally. The result of the comparative test among 131I isotope method, the “gold standard”, and the proposed method indicated that the modified pulse dye densitometry reduced the average relative error of CBV detection from 6.85% to 4.53%, which improved the detection accuracy.