The removal of inorganic arsenic (As) species from water using bone char pyrolyzed at 900 °C was investigated. Results revealed that the Sips model resulted in the best As(III) experimental data fit, while As(V) data were best represented by the Langmuir model. The adsorption rate and mechanisms of both As species were investigated using kinetic and diffusional models, respectively. At low As(III) and As(V) concentrations of 0.5 and 2.5 mg/L, the removal was due to intra-particle interactions and pore diffusion following Pseudo-first-order kinetics. However, at higher concentrations of 5 and 10 mg/L, the pore diffusion mechanism was ineffective, and the adsorption was best described by Pseudo-second-order and Elovich models. The goodness of the fit of linearized and nonlinear forms of all models against experimental data was thoroughly tested using error function analysis. Nonlinear regressions produced lower error values, so they were utilized to calculate the parameters of the models. The changes in bone char surface chemistry were examined using FTIR and Energy-dispersive X-ray spectroscopy (EDS). Arsenic oxide and complexes with metals were the confirmed immobilized forms of As on the bone-char surface. To the authors’ knowledge, this study is the first attempt at As(III) adsorption analysis using bone char. Display omitted •Removing inorganic arsenic species from water using bone char was investigated.•Arsenite adsorption onto bone char was systematically studied for the first time.•Isotherms and kinetics studies were employed to identify the nature of adsorption.•Intraparticle and pore diffusion models were used to study As-char interaction.•Initial concentration of both species determines the nature of adsorption.