Abstract: In arid or semi-arid areas, brackish water is often used for irrigation because of water shortage and soil nutrient deficiency. However, irrational irrigation with brackish water may lead to the accumulation of saline ions such as sodium, which might inhibit plant growth or even crop failure. In this study, we investigated the relationship between different irrigation methods and crop growth and yield in order to find a proper method of irrigation and fertilization. In order to achieve this objective, 2 years of field experiments in 2012 and 2013 were conducted at the experimental station in Korla, Xinjiang, China. Three irrigation treatments with brackish water, fresh water and alternative irrigation (20% of fresh water at bud and early boll stages, 80% of brackish water at late boll and boll opening stages) in the whole growing season were designed. Cotton and soil samples (from 0-100 cm depth) were collected in 4 days after the last irrigation. Cotton root were sampled with square monolith method in the space of 70cm×20cm×50cm. Root length density was estimated with DT-SCAN software. Available content of copper, iron, magnesium, calcium, potassium, and sodium in soil samples were extracted with Mehlich 3 reagent. Element content of cotton bolls were extracted after wet acid digestion (HNO3-HClO4-HF). Then, they were determined in triplicate by inductively coupled plasma optical emission spectrometer (ICP-OES). Accumulation coefficient was defined as the ratio of nutrient content in soil layers of 0-10 cm or 0-20 cm to that in 0-100 cm, representing the element aggregation capability in topsoil. The results showed that when sodium did not accumulate in 2012, irrigation with brackish water significantly enlarge (compared to results in fresh water) the aboveground dry matter, numbers of bolls per unit area, and seed cotton yields by 32%, 20%, and 22%, respectively. Nevertheless, continuous irrigation using brackish water under heavy application of nitrogen fertilizer resulted in the surface aggregation of sodium, which significantly inhibited the root growth as well as aboveground dry matter and the yields. From 2012 to 2013, the total root length density decreased significantly under the brackish water treatment, which led to a sharp reduction of the aboveground dry matter, the numbers of bolls per unit area, and the seed cotton yields (P<0.05). In this case, the root length density, the aboveground dry matter, the numbers of bolls per unit area, and the seed cotton yields were improved by 12%-24% compared to the results for the brackish water treatment under the alternative irrigation. Moreover, the alternative irrigation did not essentially solve the problem of nutrient deficiency under the traditional cultivation pattern, thus a sustainable scheme of irrigation and fertilization should be made according to the solute variation in the soil. The average copper, iron, and magnesium in the soil profile reduced by 40% from 2012-2013 and the accumulation coefficients varied from 0.14-0.17 to over 0.20; however, the average calcium, potassium, and sodium increased by 33%-45% in the soil profile. For all treatments, calcium and potassium did not gather in the topsoil, whereas sodium under the brackish water treatment changed from weak to strong aggregation in the surface. This result provides valuble information for the rational irrigation and fertilization with brackish water in cotton cultivation.