Walnut blight, caused by Xanthomonas arboricola pv. juglandis, has been responsible for significant losses in early-leafing walnut crops in Tasmania, Australia. Research over three growing years was undertaken to study disease epidemiology and refine the copper spray programme. The progression of disease incidence was best described by the monomolecular growth model in 2004-2005 and 2006-2007, implying a monocyclic epidemic. In contrast, the logistic growth model described a polycyclic epidemic in 2005-2006. For monocyclic epidemics, two sprays of copper formulated with mancozeb applied one week apart from 5% terminal bud-burst provided commercially acceptable control. For the polycyclic epidemic, four copper-mancozeb applications from bud-burst significantly reduced blight incidence in comparison to three or fewer applications; however, disease control was inadequate. Reducing active copper from 1.5 kg to 0.75 kg per hectare in monocyclic epidemics maintained control of blight. For polycyclic epidemics, half rates of copper resulted in similar levels of blight infection compared to the full rate; however, full rates significantly increased nut yield in comparison to nontreated plots. This study indicated that strategic application of copper-based bactericides, timed according to pathogen activity, may adequately control walnut blight and limit unnecessary applications of copper.