Piroplasmosis has been identified as a possible cause of mortality in reintroduced Przewalski's horses (Equus ferus przewalskii) in the Dsungarian Gobi (Mongolia). A cross-sectional and a longitudinal study were conducted in a representative sample (n = 141) of the resident domestic horse population and in 23 Przewalski's horses to assess the prevalence of Theileria equi and Babesia caballi. Piroplasms were detected in blood by light microscopy in 6.7% (95% confidence interval [CI]: 3.6-12.2%) of the domestic horse samples. Antibody prevalence was 88.6% (95% CI: 82.4-92.9%) for T. equi and 75.2% (95% CI: 67.4-81.6%) for B. caballi. Antibody prevalence did not change over time, but antibody prevalence for both piroplasms were significantly lower in animals less than 1 yr of age. For both piroplasms, the prevalence of presumably maternal antibodies (falling titers) in foals was 100%. Only one of 16 foals seroconverted against T. equi during the study period, despite that piroplasms were found in two other individuals. The incidence density (ID) of T. equi in foals was therefore 0.0012 seroconversions per horse day (95% CI: 0.00029-0.0057). In contrast, yearlings had an ID of 0.0080 (95% CI: 0.0049-0.010) for T. equi and 0.0064 (95% CI: 0.0036-0.0093) for B. caballi, and in seven individuals piroplasms were detected. The seroprevalence of both piroplasms rose from 20% in spring to 100% in autumn. Comparison of domestic and Przewalski's horses resulted in a standardized prevalence ratio (SPR) of 0.98 (95% CI: 0.80-1.24, not significant) for B. caballi; in contrast, the prevalence of T. equi in Przewalski's horses was significantly lower than expected (SPR = 0.51, 95% CI: 0.50-0.64).