Incidence-severity relationships for coffee rust were studied to determine if the easily assessed incidence could be used to evaluate host resistance and fungicide treatment. At two locations in each of 3 years, the incidence of rust on 300 leaves was compared with two assessments of severity: (i) the average number of sporulating pustules per leaf, and (ii) the estimated leaf area with rust. For nine or 10 assessments in time at one location and pooled over 3 years, the average number of sporulating pustules per leaf (Y _sp) was well related with the incidence of leaves with rust (X) as Y _sp = 0.02982+ 0.017035X +0.000573X ²; R ² = 0.87. The leaf area with rust (Y _la) was also well related with incidence of leaves with rust as Y _la = 0.001 - 0.01076X +0.008376X ²; R ₂ = 0.92. For two independent data sets from a second location obtained over two seasons, the above models satisfactorily fit the relationships for the average sporulating pustules per leaf (R ² = 0.97 and 0.96) and for the estimated leaf area with rust (R ² = 0.95 and 0.98). Therefore, the readily determined incidence can be used to estimate both measures of disease severity of coffee rust.