Inertial migration of a single cylindrical particle in a constricted microchannel is addressed in this work. A computational model (two‐dimensional) has been constructed with the assistance of the immersed boundary finite volume method. The feedback forcing strategy is utilized for the simulation of lateral migration. The parameters like equilibrium position, migration time, and shortest equilibrium distance are computed to analyze the inertial migration characteristics of the particle. Also, a comprehensive parametric study has been performed on the migration behaviour of particles inside the constricted channel by addressing the effects of Reynolds number, diameter, initial release position, and constriction clearance. The parametric study shows that the equilibrium position changes with variations in the initial release position and particle diameter. On the other hand, it stays unaffected by changes in Reynolds number and constriction clearance. The parameters like the shortest equilibrium distance and migration time increase with a rise in Reynolds number and particle diameter. On the other hand, it reduces with the reduction in constriction clearance. Inspired by the parametric study results, in the following stage, a prediction model is created with an artificial neural network algorithm. This is used for an effective forecast of equilibrium position, migration time, and shortest equilibrium distance. Further, the computational model is utilized to check for the existence of a critical Reynolds number for the particle movement in a constricted microchannel. It is observed that the critical Reynolds number remains unchanged with a change in particle diameter. However, it increases linearly with an increase in constriction clearance.