AbstractBACKGROUNDBlack leaf streak disease (BLSD) is the most important disease of bananas for export. The successful control of BLSD requires an intensive use of systemic fungicides, leading to the build‐up of resistance and failure of control. Early detection of fungicide resistance is crucial to drive rational chemical strategies. Present methods relying on ascospore germination bioassays have several drawbacks that could be overcome using conidia.RESULTSGenerally, a single genotype is present on the conidial population derived from one lesion. Conidial germination tests with thiabendazole (5 mg L−1) enable a clear detection of strains resistant to methyl benzimidazole carbamates. Germination bioassays on azoxystrobin (10 mg L−1) enable the detection of most QoI‐resistant strains, but their proportion might be underestimated with cut‐off limits of germ tube length (L > 120 µm) or growth inhibition (GI < 50%). The level of fungicide resistance differs at different canopy levels of a banana tree, which should be considered for sampling. The ascospore germination bioassay provided more variable estimations of the level of resistance by comparison with the new conidial germination bioassay.CONCLUSIONGermination bioassays performed with conidia obtained from young lesions overcome most drawbacks encountered with ascospore germination bioassays and could be considered as a new reference method for fungicide resistance monitoring in this species. Different steps are proposed, from sampling to microscopic examinations, for the implementation of this technique. © 2014 Society of Chemical Industry