Constructed wetlands are nature-based solutions able to remove different pollutants from water, including arsenic. Arsenic is a pollutant of concern given its toxicity and its presence in water sources worldwide. Despite the increased interest in investigating the performance of constructed wetlands in the treatment of arseniccontaminated water at the laboratory scale, the application of these solutions at the pilot and full scale is still limited. To understand and predict the removal of arsenic in constructed wetlands, some numerical models have been developed. Among black box models, only first-order models have been proposed, with unsuccessful results. The model that best describes arsenic retention processes in constructed wetlands is RCB-ARSENIC, a mechanistic model adapted from Retraso-CodeBright that simulates arsenic reactive transport. This model includes arsenic precipitation, arsenic sorption on supporting media, arsenic sorption on plants roots and arsenic uptake by plants; represented in the reactive term of the reactive transport equation. Thus, it includes two of the three main processes that remove arsenic in constructed wetlands: precipitation, sorption, and coprecipitation. Despite this, and what is known about arsenic geochemistry, the formulation of these reactive rates requires improvement. In addition, this model was calibrated and validated using data from a single horizontal subsurface flow constructed wetland system, which treated one type of synthetic water. Therefore, it cannot be applied to other types of arsenic-contaminated water or other constructed wetland systems. Moreover, the reactive transport of relevant species -especially iron- and their role in arsenic removal, along with relevant redox reactions associated to the presence of organic matter, oxides or bacteria-, must be included. A comprehensive mechanistic model