Abstract: Based on long-term field experiments at Dingxi in the western Loess Plateau, the effects of different agricultural measures on soil quality were studied using an integrated assessment model. A total of 6 agricultural measures with 2 tillage systems and 3 ground-cover forms were used. The agricultural measured included conventional tillage (T), no-tillage (NT), conventional tillage with straw incorporation (TS), no-tillage with straw mulching (NTS), conventional tillage with plastic-film mulch (TP) and no-tillage with plastic-film mulch (NTP). The experiment was conducted in two rotation sequences. Sequence 1 started with pea followed by spring wheat (P→W) whereas sequence 2 started with spring wheat followed by field pea (W→P). The results showed that soil quality evaluation with integrated-weighted and addition-multiplication methods reasonably reflected actual soil conditions, and the effect of tillage variations on soil quality. The 8 most sensitive indicators were selected from a total of 29 assessment indictors vial discriminative analysis. The selected indicators were used to evaluate soil physical quality, fertility, health and land productivity in the semi-arid areas of the Loess Plateau. These 8 indicators included soil aggregates, organic matter, total nitrogen, available nitrogen, invertase activity, grain yield, net income and output/input ratio. Based on the 29 indicators, the order of soil quality index (SQI) ranked from high to low was NTS, TS, NTP, NT, TP and T in both rotation sequences. Again based on the 8 indicators, the orders of SQI ranked from high to low were NTS, TS, NTP, NT, T and TP in the P→W rotation sequence, and NTS, TS, NT, NTP, T and TP in the W→P rotation sequence. The results of the 2 assessment systems were similar and highly correlated. This indicated that the 8 indicators were sufficiently representative of the soil conditions in the western Loess Plateau.