The main objective of this paper is to study the influence of wear testing parameters on the weight loss of AZ91 magnesium alloy which was developed via rotary die-equal channel angular pressing (RD-ECAP) method through the different number of passes. The AZ91 sample was pressed up to 16 passes in order to decrease the grain size to 10 μm. After RD-ECAP processing, significant grain refinement was found, which yielded a refined grain size after 16 passes and enhanced mechanical properties and wear resistance. Dry sliding wear experiments with a ball on disc apparatus at room temperature were conducted, following RD-ECAP processing. The response surface methodology (RSM) was used to optimize and compare the wear resistance of the RD-ECAP alloy and the as-received alloy. Analysis of the wear mechanism and the worn surfaces of AZ91 samples was performed using scanning electron microscope. RSM analysis showed that increased applied load, sliding speed and sliding time in all cases decreased wear resistance. On the other side, wear resistance in AZ91 magnesium alloy was improved by the increase in the number of passes of RD-ECAP. The three empirical models were validated by conducting confirmation tests and certified that the proven empirical wear models are reliable for wear applications as a predictive tool.