Synthetic media should be designed for the production of Metarhizium anisopliae conidia with improved virulence properties. A genetic algorithm (GA), demonstrated to be suitable for the design of media for spore mass production (Hutwimmer et al. 2008), was utilized for a multi-objective medium design to improve conidia yield and three proposed virulence properties of conidia: C : N ratio, germination speed and amount of spore-bound Pr1 protease. After five iterative optimizations, 52 media were improved over Sabouraud dextrose agar (SDA). Four media exhibited medium performances (a factor derived from the four single optimization variables) of around 0.7; cf. SDA = 0.532; media with enhanced properties were reached for each single optimization variable; Bioassays against Tenebrio larvae indicated also a slight improvement in virulence of conidia from designed media. A degenerated phenotype of the same strain did not exhibit differences in colony appearance, spore characteristics and virulence if grown on designed media. The application of a problem-oriented GA is a practical and rapid method to design media for multi-objective purposes. The applicability of a GA for multi-objective medium design was demonstrated for the cultivation of anamorphic fungi on solid media.