Background Accurate diagnosis of invasive mold diseases (IMD) remains challenging. Here, the performance of panfungal PCR, Aspergillus and MucoralesPCR in bronchoalveolar lavage (BAL) was evaluated. Methods We conducted a single‐center study including 167 hematologic patients at risk for IMD with BAL performed 2011‐2014. Diagnostic performance of single tests (Aspergillus‐, Mucorales‐, and panfungal PCR, galactomannan (GM)≥0.5 and ≥1, culture/cytology) or in combination was calculated for predicting IMD comparing proven/probable or proven/probable/possible IMD vs no IMD, respectively. Results IMD was classified as proven (n = 6), probable (n = 31), possible (n = 29) and no IMD (n = 101) according to EORTC/MSG criteria. GM ≥ 0.5 in BAL showed the highest sensitivity with 81% for diagnosing IMD whereas the other tests only 5%‐35%. By contrast, specificity was highest for panfungal PCR with 99% and GM ≥ 1, Mucorales and AspergillusPCR reached specificity ≥91%. When combining the tests, GM ≥ 0.5 and panfungal PCR show a sensitivity and specificity of 87% and 78% for IMD or with AspergillusPCR a sensitivity and specificity of 88% and 72% for invasive pulmonary aspergillosis, respectively. Including possible IMD patients did not improve the sensitivity of PCRs. In probable/proven IMD patients, the addition of panfungal PCR resulted further in detection of Fusarium species and Alternaria species, and the MucoralesPCR was positive in 2 probable IMD cases. Conclusion This study illustrates that the diagnosis of IMD is still very problematic and lacks objectivity. Together with GM in BAL, the PCRs may prove an addition to the current available diagnostic armamentarium in IMD because of their ability to identify molds on a species level.