Diagnostic efficacy of Galactomannan (GM) assay for invasive aspergillosis (IA) is variably reported. Data from developing countries are scant. Children with haematological malignancies and fever were enrolled prospectively. Blood sample for GM was drawn on the day of admission; levels were measured with Platellia Aspergillus enzyme immunoassay. Diagnostic criteria were adapted from EORTC-MSG-2002. Proven, probable and possible episodes were considered as the disease group. One hundred febrile episodes in 78 patients were evaluated. The mean age was 6.1 years. Majority (75%) episodes were in patients with acute lymphoblastic leukaemia. One episode each was diagnosed with proven and probable IA, while 23 were diagnosed with possible IA. Best results were obtained with a cut-off value of 1.0, with sensitivity, specificity, positive and negative predictive value of 60%, 93%, 75 and 87 respectively. The sensitivity dropped to 40%, at cut-off value of 1.5 and specificity was 38%, at a cut-off of 0.5. A higher value of GM correlated with pulmonary nodules (P = 0.037) and mortality (P = 0.001). GM assay is adjunctive to clinical/radiological evidence. A negative GM assay may not reassure the physician against the use of amphotericin in patients with febrile neutropenia, as it does not exclude the diagnosis of clinically relevant other fungal infections, particular mucormycosis.