There is limited information about the accuracy of molecular and serological diagnostic assays for tropical theileriosis in asymptomatic carrier large ruminants. This study has estimated the sensitivity (Se) and specificity (Sp) of PCR and an indirect fluorescent antibody test (IFAT) in the diagnosis of tropical theileriosis in cattle and buffaloes via a Bayesian latent class analysis (BLCA) framework. Blood samples were collected from 70 cattle and water buffaloes (Bubalus bubalis) raised under a smallholder production system in different Egyptian localities. T. annulata infection status was detected by PCR, and IFAT and the test results were subjected to BLCA without assuming the existence of a reference test. Our findings showed that the performance of PCR was superior to that of IFAT. PCR showed a higher Se [0.83 (95% PCI: 0.63–0.98)] in comparison to IFAT [0.72 (95% PCI: 0.68–0.75)]. Similarly, PCR showed a higher Sp [0.95 (95% PCI: 0.77–1.00)] than IFAT [0.82 (95% PCI: 0.80–0.84)]. Se and Sp of the two tests did not differ by species implying that the diagnostics’ performance for T. annulata infection in bovines is the same regardless of the species under consideration. In conclusion, PCR outperforms IFAT in the detection of T. annulata infection and can thus be applied to routine control of tropical theileriosis in endemic situations where cattle and buffaloes are kept under traditional smallholder production systems.