Biota could be ecosystem engineers in generating an intrinsic heterogeneous landscape through scale- dependent feedbacks. Thereby, they can form resource- enriched patchiness or islands of fertility, comprising self- organizing spatial patterns. Research so far has largely focused on the self- organized spatial patterns of plant communities in drylands. It, however, remains unclear whether and how biocrusts having distinct morphology and life history from plant communities could self- organize themselves and form unique spatial patterns. Here, we conducted field observations of biocrusts across successional stages and employed a probabilistic cellular automaton model to investigate the distinct self- organized spatial patterns exhibited by mosaic patches of mosses and lichens with different patch size distributions (PSDs). Our study demonstrates that short- range positive feedbacks initially promote the development of patches, featured with a heavy- tailed PSD, while long- range negative feedbacks subsequently curtail further expansion of big patches, thereby establishing a characteristic patch scale with regular PSDs. Strikingly, only lichens reverted back to the heavy- tailed PSD in the late succession stage, presumably implying self- organized critical fragmentation of lichen patches. Field measurements of biocrust performance at the center and edge of patches of varying sizes along succession stages further support the classic scale- dependent feedback mechanism for Turing pattern formation. Collectively, our results clearly demonstrate the capability of the biocrust communities to self- organize themselves to form distinct spatial patterns governed by the spatial and temporal scale-dependent feedbacks, potentially impacting dryland ecosystem functions and resilience.