Pneumocystis jirovecii pneumonia (PJP) in immunocompromised patients entails high mortality and requires adequate laboratory diagnosis. We compared the performance of a real time-PCR assay against the immunofluorescence assay (IFA) in the routine of a large microbiology laboratory. Different respiratory samples from HIV and non-HIV-infected patients were included. The retrospective analysis used data from September 2015 to April 2018, which included all samples for which a P. jirovecii test was requested. A total of 299 respiratory samples were tested (bronchoalveolar lavage fluid (n = 181), tracheal aspirate (n = 53) and sputum (n = 65)). Forty-eight (16.1%) patients fulfilled the criteria for PJP. Five positive samples (10%) had only colonization. The PCR test was found to have a sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of 96%, 98%, 90% and 99%, compared to 27%, 100%, 100% and 87%, for the IFA, respectively. PJ-PCR sensitivity and specificity were >80% and >90% for all tested respiratory samples. Median cycle threshold values in definite PJP cases were 30 versus 37 in colonized cases (p < 0.05). Thus, the PCR assay is a robust and reliable test for the diagnosis PJP in all respiratory sample types. Ct values of ≥36 could help to exclude PJP diagnosis.