Paddy soils of over 500 hectares had been polluted by arsenic (As) from tailings at an abandoned lead‐zinc mine at Shaoxing, Zhejiang, China. Several field experiments were conducted to establish measures for reducing As toxicity to rice plants. The results obtained were as follows. Fresh Chinese milkvetch (Astragalus sinicusL.) was not supposed to be used as green manure in arsenic polluted paddy soils. Although liming (1,500 kg CaO hectare‐1) could reduce water‐soluble As (H2O‐As) in the soil, the rice plant grew badly. The treatments of FeCl3(25 mg Fe kg‐1soil) and MnO2(25 mg Mn kg‐1soil) could markedly lower the H2O‐As and arsenite [As(III)] percentage in the soil and make the plant grow better than the control experiment (CK). Without adding any materials to the soil, wetting and drying (furrowing and draining) in the paddy soil could increase soil redox potential greatly and lower the H2O‐As and As(III) percentage obviously leading to better rice growth. In addition, the As contents of roots, flag leaf, grain, and husked rice of 11 new cultivare of early rice were determined and correlation analysis was conducted. Uptake and accumulation of As in different parts of cultivars Zhefu‐802 and Erjiufeng at the 4 As levels of the paddy soil demonstrated that the As contents in husked rice of both cultivars exceeded the hygienic standard (0.7 mg As kg‐1) when they grew in the paddy soil having total As content of about 70 mg kg‐1for Zhefu‐802 and 100 mg kg‐1for Erjiufeng, respectively.