1. Dead wood is a habitat for numerous fungal species, many of which are important agents of decomposition. Previous studies suggested that wood-inhabiting fungal communities are affected by climate, availability of dead wood in the surrounding landscape and characteristics of the colonized dead-wood object (e.g. host tree species). These findings indicate that different filters structure fungal communities at different scales, but how these factors individually drive fungal fruiting diversity on dead-wood objects is unknown. 2. We conducted an orthogonal experiment comprising 180 plots (0.1 ha) in a random block design and measured fungal fruit body richness and community composition on 720 dead-wood objects over the first four years of succession. The experiment allowed us to disentangle the effects of the host (beech and fir; logs and branches) and the environment (microclimate: sunny and shady plots; local dead wood: amount and heterogeneity of dead wood added to plot). 3. Variance partitioning revealed that the host was more important than the environment for the diversity of wood-inhabiting fungi. A more detailed model revealed that host tree species had the highest independent effect on richness and community composition of fruiting species of fungi. Host size had significant but low independent effects on richness and community composition of fruiting species. Canopy openness significantly affected the community composition of fruiting species. By contrast, neither local amount nor heterogeneity of dead wood significantly affected the fungal diversity measures. 4. Synthesis: Our study identified host tree species as a more important driver of the diversity of wood-inhabiting fungi than the environment, which suggests a host-centred filter of this diversity in the early phase of the decomposition process. For the conservation of wood-inhabiting fungi, a high variety of host species in various microclimates is more important than the availability of dead wood at the stand level.