Acid mine drainage (AMD) commonly contains elevated concentrations of As(III) and/or As(V) due to oxidation of arsenic-containing sulfides. Bone char has been used as a low-cost filling material for passive treatment. The breakthrough curves of As(III) and As(V) were studied in column experiments conducted at different flow rates, adsorption cycle times, and with different coexisting cations and anions to compare their transport behaviours. The experimental data were fitted by the Convection-Diffusion Equation (CDE) and Thomas model with the aim of obtaining retardation factors of As(III) and As(V) and their maximum adsorption capacities, respectively. The maximum adsorption capacities of As(III) and As(V) are 0.214 and 0.335 mg/g, respectively. Coexisting Mn2+ and Al3+ ions can shorten the equilibrium time of As(V) adsorption from 25 h to 8 h, but they have little effect on As(III). The retardation factors of As(III) and As(V) calculated by the CDE model decrease with adsorption cycles from 37 to 20 and 51 to 32, respectively. The Mn2+ and Al3+ ions could enhance retention ability with adsorption cycle time, especially Mn2+ for As(V). Secondary adsorption phenomena were observed only in breakthrough curves of As(V) in the presence of Mn2+ and Al3+. The competitive influences of coexisting arsenic species is As(V) > As(III). Regeneration experiments using distilled water and NaOH solution were completed to quantify the degree of desorption of both As(III) and As(V). The results show that As(V) adsorbed on bone char has better desorption performance than As(III), and the average degrees of desorption of As(III) and As(V) for three desorption experiments are 75% and 31%, respectively.