The substitution of conventional urea with controlled-release fertilizers (CRFs) offers a more efficient fertilization strategy that reduces chemical nitrogen input while enhancing crop yield and quality. A field experiment was conducted from 2022 to 2023 in Gaoping town, Jingchuan county, located on the arid plateau region of the Longdong Loess Plateau, using the sorghum variety Jiza 127. Seven treatments were implemented: no nitrogen application (CK), conventional urea at 240 kg hm-2 of pure nitrogen (CU240), and five CRF-based treatments with varying nitrogen reduction rates—0 (CRU240), 10% (CRU216), 20% (CRU192), 30% (CRU168), and 40% (CRU144). The study aimed to evaluate the effects of replacing conventional urea with CRFs on dry matter accumulation and partitioning, as well as yield and grain quality in sorghum. Results showed that throughout the growth period, the dry matter accumulation rate (DMAR) followed a typical pattern of increasing and then decreasing, peaking during the jointing-to-heading stage. Logistic models fitted to DMA dynamics demonstrated strong goodness-of-fit (R2 = 0.949-0.985) across all treatments. Nitrogen application reduced the proportion of dry matter retained in stems and leaves but significantly enhanced translocation to the panicle at maturity. Notably, the proportion of dry matter allocated to the panicle in CRU240 and CRU216 treatments was, on average, 1.78% and 0.94% higher, respectively, than in CU240. Before flowering, pre-flowering dry matter translocation (DMR), translocation efficiency (DMRE), and the contribution of pre-flowering translocation to grain dry matter accumulation (DMRCG) were significantly lower in CRU168 and CRU144 compared to CU240, while no significant differences were observed in CRU240, CRU216, and CRU192. After flowering, post-flowering DMA in CRU240, CRU216, and CRU192 treatments was comparable to CU240, but DMA in CRU192 was 8.86% and 14.15% higher than that in CRU168 and CRU144, respectively. The highest yield was achieved under CRU240, which was significantly greater than CRU168 and CRU144 by 19.36% and 33.01%, respectively. As nitrogen input decreased, yield declined accordingly, with reductions of 3.63%, 8.83%, 16.20%, and 24.82% observed in CRU216, CRU192, CRU168, and CRU144, respectively, compared to CRU240. Grain protein content peaked in CRU240, being significantly higher than in CRU168 and CRU144 by 19.92% and 20.70%, respectively. In contrast, starch content increased as nitrogen input decreased. Tannin content did not differ significantly among treatments. Path analysis indicated that panicle grain weight and 1000-grain weight were the key determinants of high and stable yields. In conclusion, compared to conventional urea, reducing controlled-release nitrogen application by 10%-20% effectively promotes dry matter accumulation, partitioning, and translocation, while maintaining high yield and improving grain quality. This strategy enhances both agronomic performance and resource use efficiency and is therefore recommended as an optimal nitrogen management approach for dryland sorghum production.