Objective To establish a high-throughput screening system for anti-Babesia compounds to identify highly active candidate compounds, and to evaluate the in vivo efficacy of these compounds against Babesia infections in animal models. Methods Specific primers were designed targeting B. duncani histone-like transcription factor gene (histf; GenBank accession number: XM_067945812). A B. duncani culture with an initial parasitemia of 21.48% was subjected to serial dilutions to prepare 7 samples with different parasitemia levels (2.148 × 10^-1 to 1.375 × 10^-5). Genomic DNA was extracted from each sample, and the histf gene was amplified by quantitative Real-time PCR (qPCR) assay to construct a standard curve correlating parasite burden with cycle threshold (Ct) values. B. duncani-infected erythrocytes were seeded onto 96-well plates (5.5 × 10⁴ cells/well), followed by addition of 1 212 compounds from the Anti-Infection Compound Library at a final concentration of 1 × 10^-5 mol/L to evaluate the anti-parasitic activity. The 20 most active compounds were serially diluted (1 × 10^-7 to 3 × 10^-5 mol/L) and incubated for 48 h and 72 h, respectively. Genomic DNA was then extracted from parasites in each group for amplification of the histf gene using qPCR assay to yield Ct values, and the inhibitory rates and in vitro half-maximal inhibitory concentrations (IC₅₀) were calculated. HEK-293T cells were seeded onto 96-well plates at 8 000 cells per well. The 20 most active compounds were serially diluted into concentrations of 1 × 10^-6, 3 × 10^-6, 1 × 10^-5, 3 × 10^-5, 1 × 10^-4, and 3 × 10^-4 mol/L and added to respective wells, and a vehicle control group (without compound treatment) was also assigned. Cell viability was assessed using the CCK-8 assay by measuring the absorbance at 450 nm (A₄₅₀ value), and the half-maximal cytotoxic concentration (CC₅₀) were calculated. Twenty golden hamsters and 12 BALB/c mice were intraperitoneally inoculated with B. duncani-infected erythrocytes (1 × 10² cells/animal) and B. microti-infected erythrocytes (1 × 10⁴ cells/animal), and then randomly assigned to control and treatment groups. Animals in treatment groups were intraperitoneally injected with dacinostat at a dose of 10 mg/kg 2 days post-infection for 15 consecutive days, while animals in controls groups were given the same volume of vehicles. Giemsa-stained blood smears were prepared daily, and erythrocyte numbers were counted and parasitemia was determined to evaluate in vivo efficacy. All statistical analyses were performed using the software GraphPad Prism 10.1.2, and differences of means between groups were tested for statistical significance with unpaired t test. Results qPCR amplification showed normal amplification curves using histf as the target gene, and the melting curve exhibited a single peak, indicating a stable qPCR system and highly specific amplification products. Ct values showed a good linear correlation with parasitemia, with a standard curve slope of ‒3.734, a coefficient of determination (R²) of 0.993, and a regression equation of y = ‒3.734x + 32.272. Standard curve analysis showed that Ct values presented a tendency towards a rise with a reduction in parasitemia. Screening of 1 212 compounds from the compound library showed inhibitory rates of ‒47.1% to 98.1% against B. duncani, including 31 compounds with inhibitory rates of > 80%, and 20 compounds showed in vitro inhibitory activity against B. duncani, with IC₅₀ values ranging from 3.3 × 10^-8 to 1.325 × 10^-5 mol/L. Dacinostat showed the strongest inhibitory activity, with an IC₅₀ of 3.3 × 10^-8 mol/L. The cytotoxicity of each compound against HEK-293T cells increased over time, and six compounds, including dacinostat, camostat mesylate and fenchol, exhibited both high anti-parasitic activity and low cytotoxicity. In golden hamster models of B. duncani infections, dacinostat delayed parasite detection and significantly reduced parasitemia relative to vehicles 12 days post-infection [(1.75 ± 0.94)% vs. (17.24 ± 3.15)%; t = 4.798, P < 0.05]. Two animals survived at the end of the experiment, while all animals died in the control group. In BALB/c mouse models of B. microti infections, dacinostat also significantly reduced parasitemia relative to vehicles 15 days post-infection [(0.06 ± 0.02)% vs. (6.99 ± 1.19)%; t = 5.813, P < 0.05], and parasitemia remained at a low level ranging from 0.9% to 1.0% prior to 23 days post-infection. Conclusion A high-throughput screening approach for anti-Babesia drugs has been successfully established, and multiple compounds with high in vitro anti-parasitic activities have been identified. Dacinostat shows favorable efficacy against both B. duncani and B. microti.