When we simulated backscattering coefficients (b(bp)) using quasi-analytical algorithm and optical closure theory, it was found that the position of reference wavelength (lambda(0)) had great impact on final results. In order to identify influence factors of the optimum lambda(0), datasets of Lake Taihu, Lake Chaohu and Lake Dianchi were used. After simulating of b(bp), the relationship between optimum lambda(0) and associated water quality parameters were analyzed. Meanwhile, power functions were utilized to model b(bp) spectra, and spectral slope parameters of Lake Taihu, Lake Chaohu and Lake Dianchi were 2.643 +/- 0.317, 2.719 +/- 0.242, 1.638 +/- 0.534, respectively. The results indicate that: (1) When the whole lakes are taken as objects of study, the lambda(0) should be changed to longer wavelength with the increasing of total suspended particle matter concentration (c(SPM)), suspended organic particle matter concentration (c(SPOM)) and chlorophyll a concentration (c(CHL)), the optimum location of lambda(0) of Lake Taihu is 695 nm, Lake Chaohu is 720 nm and Lake Dianchi is 730 nm; (2) When the samples in lakes are taken as objects of study, the influence factors are different due to the complexity of optical properties of inland waters. But it is to be noted that, c(SPM) is a key factor in common. Besides, the regularity mentioned above should be weakened with high c(CHL). (3) The relationship between b(bp) and c(SPM) is much stronger in the waters with suspended particles dominated by inorganic particles, in Lake Taihu, for example, relationship between these two variables is strong with R2 = 0.852.