Abstract: The semiarid western Loess Plateau is characterized by hilly landscape that is severely prone to soil erosion. Stability and distribution of soil water-stable aggregates could be affected by soil tillage methods in dry land areas. An improved understanding of the effect on soil and water erosion associated with the production of land is required for enhancement of agricultural sustainability in semiarid areas. A 15-year local field experiment was carried out to study the effects of different tillage methods and straw applications on soil water-stable aggregates and aggregate destruction mechanisms under spring-wheat/pea rotation using three Le Bissonnais (LB) and routine wet sieving (RW) methods. Three LB wet sieving methods used in the experiment were slow wetting sieving (SW) method simulating light rains (micro-cracking), fast-wetting sieving (FW) method simulating heavy rains (slaking), and wet stirring sieving (WS) method simulating disturbance (mechanical breakdown). Four aggregate size ranges were obtained by the sieving methods:2-5 mm (larger aggregate, LA); 0.25-2 mm (small aggregate, SA); 0.053-0.25 mm (micro-aggregate, MA); and < 0.053 mm (slit plus clay, SC). The results of the three LB methods and RW method were then compared and the mean weight diameter (MWD), relative slaking index (RSI) and relative mechanical breakdown index (RMI) of soil aggregates were calculated. The field experiment was located in the Rainfed Agricultural Experimental Station (35°28'N, 104°44'E) which belongs to Gansu Agricultural University in Dingxi, Gansu Province, China. The experimental work included the following tillage and straw mulching treatments:conventional tillage (T), no tillage (NT), no tillage with straw mulching (NTS), and conventional tillage with straw incorporation (TS). The treatments were arranged in a complete randomized block design with three replications. The soil samples were taken at three soil depths (0-5 cm, 5-10 cm and 10-30 cm). The results showed that after wetting treatments except the slow wetting method, the dominant fraction of fragments in each soil layer was < 0.25 mm under all four tillage treatments. The order of sieving method as for < 0.25 mm non water-stable aggregates content was RW > FW > WS > SW. MWD of soil aggregates for four sieving methods was in the order of SW > WS > FW > RW under all the four tillage treatments. This trend indicated that aggregate breakdown mechanism was in the order of:slaking > mechanical breakdown > micro-cracking. While NTS treated soils exhibited the highest MWD and water-stable aggregates content for all wet sieving methods in the 0-5 cm and 5-10 cm soil layers. MWD for NTS treatment was significantly greater (P ≤ 5%) than T and NT treatments. Also TS treatment showed the highest MWD and water-stage aggregates content in the 10-30 cm soil layer, but with no significant difference in MWD from NTS. Compared with T treatment, TS treated soils significantly improved MWD. RSI and RMI of soil aggregates were suppressed by NTS, TS and NT treatments, and NTS treatment had the highest effect. Straw addition significantly suppressed RSI and RMI of soil aggregates in all three observed soil layers. No tillage significantly suppressed RSI of soil aggregate in the 0-5 cm soil layer. The results suggested that heavy rain was the main factor destroying soil aggregates in rainfed farmlands in the Loess Plateau region in Central Gansu Province. NTS treatment performed best for sustainable agricultural development and soil and water conservation in the Loess Plateau region in Central Gansu Province.