Statistical experimental design was utilized to optimize removal of aqueous Hg(II) by Fraxinus tree leaves through a batch biosorption process. Sorbent-sorbate behavior was evaluated by fitting equilibrium data by nonlinear and transformed linear forms of the Langmuir, Freundlich, and Redlich-Peterson isotherms. The comparative study showed that nonlinear regression is a better way to model equilibrium data. A 2(3) full factorial design was used to identify significant factors and interactions. The pH, Hg(II) initial concentration, and sorbent mass were examined as major factors. The contact time was fixed at 30 min. All of the factors were significant at the 95 % confidence level. The amount of Hg(II) was determined by cold vapor atomic absorption spectrophotometry. A regression model was derived by using a response surface methodology through performing central composite design (CCD). Model adequacy was checked by such diagnostic tests as analysis of variance (ANOVA), lack of fit test, residuals distribution, and over-fitting test. On the other hand a residuals distribution was evaluated for normality according to the Ryan-Joiner test. As a result, the optimized condition for Hg(II) biosorption was calculated to be pH = 4.4, s = 0.25 g, and at = 50 mg.L(-1), which corresponds to 92.25 % removal efficiency. The biosorption process was kinetically fast and followed a pseudosecond order kinetic model. Fourier transform infrared (FT-IR) and X-ray diffraction (XRD) spectra were used to find more about the biosorption mechanism.