Phosphorus (P) adsorption capacities of materials derived from batch experiments can vary by several orders of magnitude depending on the method used, leading to potential misinterpretation of the P retention capacity on a long-term basis and unrealistic estimations of constructed wetland systems (CWS) longevity. The objective of this study was to determine if the P saturation of the material in a column could be used for this purpose with an improved accuracy. A 278-d column experiment with a synthetic P solution was conducted to investigate the long-term P retention capacity of electric arc furnace (EAF) steel slag up to its P saturation point. EAF slag showed a high affinity for P, reaching a saturation value of 1.35 g of P kg(-1). Investigations of the regeneration of the P adsorbing capacity by this material showed that, after 4 weeks of water desaturated resting, EAF steel slag was able to increase its initial P adsorptive capacity to 2.35 g of P kg(-1). A sequential P fractionation experiment was performed to quantify the proportion of P bound to mineral compounds in EAF. From the most loosely bound to the most strongly bound P fraction, P was associated with resin extractable (14%), Fe extractable (0.5 M Na2CO3, 47%), Al extractable (0.1 M NaOH, 1%), Ca extractable (1 M HCl, 12%), and Ca in a stable residual pool (concentrated hot HCl, 26.5%). X-ray fluorescence analyses of EAF steel slag chemical composition revealed that the continuous application of a P solution resulted in 75% and 59% increases in K2O and P2O5 respectively; Al2O3 and FeO increased by 8%, while the portion of CaO remained unchanged. The investigated properties (P retention potential, regeneration of P adsorption, P fractionation) provide useful data about the suitability of slag material as a media for long-term P removal and should enable an improved prediction of the longevity of full-scale CWS.