The prevalence and incidence of Trypanosoma evansi infections in village buffaloes in Central Java were estimated using parasitological tests, two antigen-detection ELISAs (2G6 Ag-ELISA and Tr7 Ag-ELISA), an antibody-detection ELISA (IgG ELISA) and a card agglutination test (CATT). Of 2387 village buffaloes tested in five districts, 4% (95% confidence interval [CI]: 3%, 5%) were positive with the microhaematocrit test (MHCT), 58% (95% CI: 56%, 60%) were positive with the 2G6 Ag-ELISA and 70% (95% CI: 68%, 72%) were positive with the Tr7 Ag-ELISA. An increasing prevalence with age was found and the proportion of positive buffaloes was highest in the over 84 months-old age-group (68%) with the 2G6 Ag-ELISA and in the 37–60 months-old age-group (78%) with the Tr7 Ag-ELISA. Parasitaemic buffaloes were found in more than half of the villages visited. Corrected village-specific prevalence values obtained with the two Ag-ELISAs ranged from 0% to over 100%, and prevalence differed significantly (P[les ]0·0001) between villages in four of the five districts. Overall, 10% of buffaloes tested in markets were found to be parasitaemic and 39, 56 and 47% were found positive with the 2G6 Ag-ELISA, IgG ELISA and CATT, respectively. Incidence rates varied according to the test used and ranged from 0·22 (95% CI: 0·09, 0·44) to 0·44 (95% CI: 0·24, 0·76), per animal-year at risk, in two villages. The results highlight the importance of using validated diagnostic tests to obtain accurate estimates of prevalence and incidence. These parameters are needed, for example in mathematical models, for the development and evaluation of different control strategies for T. evansi infections in buffaloes.