Anthracnose of cashew (Anacardium occidentale) was studies on various genotypes and locations in Mozambique. Colletotrichum gloeosporioides was identified as the anthracnose causal agent using polymerase chain reaction. The relationships between incidence and severity of anthracnose on cashew genotypes were statistically analyzed by regression. Anthracnose leaf incidence, which is practically easy to evaluate, was consistently associated with leaf severity, and their relationships can be estimated using the restricted exponential function across locations, crop seasons, genotype and fungicide trials. Pooled data enabled estimation of initial incidence of 1.43% with percentage variance accounting for 83.2 and standard error of 8.3. By computing incidence data into the summary equation, 24 changes of 0, 1, 5, 10 and 40%, resulted in changes of severity estimates of 0.01, 0.05, 0.10, 0.50 and 1.00%, respectively. The maximum disease incidence was estimated as 80% when the severity reached only 5%. Increase in severity was observed afterward, approached a maximum of 25% when leaf detachment is observed. The use of incidence data for epidemic comparisons, genotype and fungicide evaluation in cashew orchards is recommended. Anthracnose incidence on leaves however, could not predict incidence on nuts.