Human infections by zoonotic encephalitis viruses are usually asymptomatic or symptoms are not specific to these viruses. Some of them have high mortality and morbidity rates and most often no specific treatment exist. This emphasizes the need for a precise identification of arboviruses in clinical specimens from humans and animals. Because these diseases are frequent in developing countries and tend to emerge or re-emerge in others, diagnostic tools must detect the broadest possible range of viruses with a high sensitivity and this is a key factor for surveillance, control of transmission and prevention through vaccination. In countries with limited diagnostic infrastructures, low-cost and easy-to-use tests are required. The diagnosis of arboviral encephalitis has been significantly improved in the recent years. Sensitive ELISA assay to detect antibodies against many arboviruses in serum or CSF are commercially available and can be used to detect early infections. Immunochromatographic rapid tests for the detection of specific IgM that could be used on fingertip blood would be valuable tools in developing countries. A limitation of these serologic assays is their lack of specificity as many arboviruses are antigenically related. Virus isolation or molecular assays from different human or animal tissues are also important diagnostic tools. Molecular assays have been extensively described in the recent years. They are very sensitive and have the advantage over cell culture that specimen transportation is less critical. Real-time detection has even improved sensitivity and reduced time-to-result. Although the utility of molecular assays for the detection of arboviruses in mosquito pools has been demonstrated, an extensive validation of their pertinence in clinical settings is still required. The use of DNA-microarrays may further extend the range of viruses that can be detected in a single test and allow isolates typing for epidemiology purposes.