Fungal infections represent a major source of morbidity and mortality in patients with almost all types of immunodeficiencies. These infections may be nosocomial (aspergillosis) or community acquired (cryptococcosis), or both (candidiasis). Endemic mycoses such as histoplasmosis, coccidioidomycosis, and penicilliosis may infect many immunocompromised hosts in some geographic areas and thereby create major public health problems. With the wide availability of oral azoles, antifungal prophylactic strategies have been extensively developed. However, only a few well-designed studies involving strict criteria have been performed, mostly in patients with hematological malignancies or AIDS. In these situations, the best dose and duration of administration of the antifungal drug often remain to be determined. In high-risk neutropenic or bone marrow transplant patients, fluconazole is effective for the prevention of superficial and/or systemic candidal infections but is not always able to prolong overall survival and potentially selects less susceptible or resistant Candida spp. Primary prophylaxis against aspergillosis remains investigative. At present, no standard general recommendation for primary antifungal prophylaxis can be proposed for AIDS patients or transplant recipients. However, for persistently immunocompromised patients who previously experienced a noncandidal systemic fungal infection, prolonged suppressive antifungal therapy is often indicated to prevent a relapse. Better strategies for controlling immune deficiencies should also help to avoid some potentially life-threatening deep mycoses. When prescribing antifungal prophylaxis, physicians should be aware of the potential emergence of resistant strains, drug-drug interactions, and the cost. Well-designed, randomized, multicenter clinical trials in high-risk immunocompromised hosts are urgently needed to better define how to prevent severe invasive mycoses.