High intensity dry magnetic separators are gaining popularity for the separation of para-magnetic minerals due to the cost economic factor. Induced roll magnetic separator is found to be an effective dry separator for the separation of fine particles. Separation efficiency of this separator depends on mineral characteristics and the design features of equipment along with the optimization of process variables. Present investigation focuses on the prediction and validation of separation performance of minerals while treating in induced roll magnetic separator. Prediction of the separation is expressed in terms of separation angle at which a particle leaves the rotor surface by using a modified particle flow model derived by Cakir. The validation of the model is carried by capturing the particle trajectory using an image analyzer. It is found that Cakir’s mathematical model produces reliable results and a new model is proposed to increase the reliability of separation angle prediction by including the particle shape factor.