In this work, the removal of Congo red dye from its aqueous solution by a low-cost adsorbent obtained from Perna viridis sea shell, an aquacultural solid waste, was studied. Maximum removal of 98% of the dye was obtained with adsorbent dosage 22 g L−1, pH of 5.5, contact time of 30 min and temperature of 308.5 K. Various adsorption reaction models were investigated and found that the pseudo-second-order model best described the reaction kinetics. Further, the mechanism of adsorption was studied using the intra-particle diffusion model, and it was confirmed that the intra-particle diffusion and boundary layer diffusion were the rate-controlling factors. Artificial neural network model was developed for the prediction of the percentage removal of the dye from its aqueous solution, and the prediction was in agreement with the experimental results obtained. It was inferred that the adsorbent prepared from P. viridis having high thermal stability (600 °C), a surface area of 13.746 m2 g−1 and mean pore diameter 7.0625 nm should be an excellent alternative for the conventional adsorbent used for the removal of the carcinogenic dye such as Congo red from its aqueous solution at any operating condition.