As the largest carbon pool in terrestrial ecosystems, soil carbon storage is fundamentally governed by microbial carbon use efficiency (CUE). Although ultraviolet (UV)-driven photodecomposition is a key pathway for the transfer of nutrients from litter to soil, its effects on microbial CUE following litter addition remain poorly understood. We conducted a two-year field experiment in an arid land to examine the effects of UV radiation on microbial CUE and community properties across two distinct litter types, and to explore the relationship between them. Results suggested that UV radiation significantly increased microbial CUE (+44.74%). The high-quality litter (C/N = 28.12) exhibited higher CUE (+99.65%) than in low-quality litter (C/N = 110.50) in UV pass condition. UV radiation enhanced bacterial biomass (+22.08%) by accelerating litter nutrients releasing and favoring the abundance of some oligotrophic taxa due to their high UV resistance. High-quality litter supports higher bacterial and fungal PLFAs and diversities than low-quality litter by providing more available nutrients, especially under UV-block conditions. Among these various microbial properties, the fungal community beta-diversity and specific functional modules within the microbial network along with extracellular enzymatic activity (EEA) imbalance emerged as key predictors of nutrient limitation. Furthermore, the microbial CUE in this region was primarily influenced by P limitation. This limitation was reduced by mitigating EEA imbalance under UV pass and high-quality litter conditions. Importantly, under UV-Pass conditions, the increasing abundance of oligotrophic taxa (particularly within specific fungal function groups) can also reduce P limitation and consequently increase microbial CUE, which may be due to their low nutrient requirements and high C use ability. Overall, our results highlight the role of litter photodecomposition in shaping microbial metabolism in arid ecosystems and suggest that incorporating shifts in microbial communities, especially specific functional groups, into models could enhance predictions of soil nutrient storage.