This study investigated the single and multicomponent adsorption of three emerging pollutants, the basic dyes Rhodamine 6G (R6G), Auramine-O (AO), and Brilliant Green (BG) by using hydroxyapatite synthesized from Pirarucu scales as adsorbent (HAP). The adsorption process was studied using seven different systems: AO-single, R6G-single, BG-single, R6G + AO, BG + AO, BG + R6G, and R6G + AO + BG. For kinetics, the initial concentration of each adsorbate per system was 50 mg/L, the results showed that the singular adsorption of these dyes was best-represented by the pseudo-second-order model (qAO = 62.54 mg/g, qR6G = 7.91 mg/g, qBG = 62.40 mg/g), however, the multicomponent adsorption was well-fitted by a pseudo-first-order model (ternary system: qAO = 56.21 mg/g, qR6G = 14.95 mg/g, qBG = 60.62 mg/g). For equilibrium, the initial concentration of each adsorbate per system was 10-300 mg/L, and the single adsorption systems were best represented by the Langmuir model. Nonetheless, the results displayed in the multicomponent mixture showed the presence of inflection points of AO and R6G whenever BG was present in solution with C0 > 150 mg/L, thus indicating that BG has greater affinity with HAP. The presence of inflection points in the curves represented a limitation for applying traditional equilibrium models, thus, an artificial neural network (ANN) was applied to non-linear curve fit this process and satisfactorily predicted the kinetics and equilibrium data. Finally, the analysis of thermodynamics for the ternary mixture revealed that the adsorption process is spontaneous (ΔG < 0), endothermic (ΔH > 0), and increases to a disorganized state as the temperature rises (ΔS > 0).