To evaluate a modified rapid ELISA method for detecting CAg during Toxoplasma gondii infection, we analyzed the specificity and sensitivity of the ELISA method by using experimental Toxoplasma infection in rabbits and also tested this method in human samples including 5428 serum, 548 cerebrospinal fluid and two breast milk samples. We prepared PcAb, and used it for rapid one-step sandwich ELISA testing in which an incubation time in the regular ELISA procedure was omitted. This method detected CAg at the concentration of 31.2 ng/mL, and no cross-reaction was found with antigens of protozoa (Cryptosporidium parvum, Plasmodium falciparum), trematode (Schistosoma japonicum, Paragonimus sp.) and nematode (Brugia malayi, Ancylostoma duodenale, Ascaris lumbricoides and Trichinella spiralis). CAg was detected in rabbit serum 3 days after infection, and optical density values reached a peak 9-13 days after infection, then declined gradually. Among human serum samples, the positive rate of CAg was 2.11% in cerebral paralysis patients, whereas it was 0.22% or 0.71% in patients without neurological symptoms or in uncomplicated pregnant women. The difference among these three groups was statistically significant (P < 0.05). The positive rate of cerebrospinal fluid samples from cerebral paralysis patients was 10.58%. There is a statistically significant difference between the positive rates of meat-packing workers and blood donors (P < 0.01). In the retrospective analysis, CAg was detected in accordance with the onset of clinical symptoms, suggesting that CAg could reflect the clinical course in humans. Together with these results, CAg detected in the modified rapid sandwich ELISA could be a sensitive marker for acute and active infection of T. gondii.