The efficacy of antifungal prophylaxis for prevention of invasive aspergillosis (IA) may depend on whether IA results from recent inhalation of spores or reactivation of latent colonisation. Compare the efficacy of liposomal amphotericin B (LAmB) for prophylaxis in acute and reactivation models of IA. In the acute model, mice immunosuppressed from day 0 were challenged at day 3 with an aerosol of Aspergillus fumigatus. LAmB (15 mg kg−1) was administered at day 0 or at challenge. In the reactivation model, naïve mice exposed to A. fumigatus remained untreated until clearance of spores from the lungs, then immunosuppressed to induce reactivation. A single LAmB dose was administered at start of immunosuppression. In the acute model, a single administration of LAmB at start of immunosuppression was not effective, but an additional administration resulted in a significant decrease in lung fungal burden (P < 0.05 vs. controls). A significant prophylactic efficacy was observed when LAmB was administered once at challenge (P < 0.01). In the reactivation model, a single LAmB administration at start of immunosuppression significantly reduced both reactivation rate and fungal burden vs. controls (P < 0.01). Our results show that the conditions under which IA develop and timing of administration of LAmB were determinant variables for prophylactic efficacy.