Direct demonstration of bacterial and/or fungal nucleic acids in the clinical samples of patients with blood stream infections is crucial in terms of rapid diagnosis, early and accurate therapy and patient management. This study was aimed to determine the presence of bacteria and fungi by polymerase chain reaction (PCR) in the clinical samples of experimental sepsis induced animals, to compare the results with culture and to evaluate the efficiency of PCR in the discrimination of bacteremia and fungemia. A total of 12 rabbits experimentally infected with standard strains of Staphylococcus aureus, Escherichia coli, Aspergillus fumigatus and Candida albicans to generate bacteremia (n= 4), fungemia (n= 4) and polymicrobial blood stream infection (n= 4), were included in the study. A total of 63 specimens of which 27 were blood and 36 were tissue (12 spleen, 12 liver, 12 kidney) samples were collected at 24, 48, 72 and 96th hours of infection. Uninfected healthy rabbits (n= 4), colony suspensions of standard bacterial and fungal strains (n= 15) and human blood samples contaminated with standard bacterial and fungal strains (n= 10) were used as controls. Microbial DNAs were searched by using real-time PCR in all the samples, and quantitative cultures were performed simultaneously. Gram-positive and gram-negative PCR protocols were performed for the samples of bacteremic animals, whereas panfungal PCR, Aspergillus and Candida PCR protocols were performed for the samples of animals with fungemia. All of those PCR protocols were applied separately for the samples of polymicrobial blood stream infection cases. Culture positivity was detected in 8 (29.6%) of the blood samples and bacterial and/or fungal DNAs were demonstrated in 20 (74%) of the blood samples by PCR. Microbial DNAs were also detected in 32 (89%) of 36 tissue samples (11 spleen, 11 liver, 10 kidney). Sensitivity rates of culture method to detect bacteremia and fungemia were 30% and 21.7%, respectively, whereas those rates were 69.2% and 69.5% for PCR, respectively. All PCR protocols gave positive results 16.5 hours before the blood cultures. Amplification of DNA from colony suspensions yielded positive results for all of the samples with a lower detection limit ranged between 30-50 cfu/ml. Detection limit of all PCR applications was 50 cfu/ml for simulated blood samples. It was concluded that the adaptation of the tested PCR method to the clinical samples obtained from patients might help to the early diagnosis of blood stream infections. However, further clinical studies are necessary to support the results of this study.