Each year, nearly two million Muslim pilgrims from all over the world come to Mecca to perform Hajj, which is the fifth pillar of Islam. The number of pilgrims increases every year, raising concerns regarding how to manage such huge crowds and how to prevent harmful incidents or crowd congestion disasters. The authorities have taken several steps to maintain free-flowing crowds; however, they continue to face difficulties. While numerous studies have been conducted to promote effective and efficient crowd management in Mecca, almost none of them have focused on crowd movements during the Rami AlJamarat ritual, which is when most of the causalities have occurred over the years. None of the studies proposed solutions to directly control the crowd movement without changing the path’s current structure. This study proposes a solution to prevent congestion at the Rami Al-Jamarat path without the need to change the structure, and it uses live crowd images to determine proper control signals that are then displayed directly to pilgrims as coloured lights. To demonstrate the results of this solution, a simulation is presented using an agent-based modelling platform.