In recent years, brownfield sites have become an important source of land for urban redevelopment in major Chinese cities. According to Chinese laws, contaminated sites have to be cleaned up before land transaction and redevelopment. National and regional technical guidelines are available to evaluate the performance of remediation. However, current guidelines are based on discrete sampling method. Due to the inherent heterogeneity of pollutant distribution, the representativeness of discrete samples has been proven questionable. In addition, relevant Chinese sampling guidance does not properly evaluate sampling error, which is the major source of error for soil environmental monitoring data. To improve sample representativeness and confidence in analytical data, decision unit multi increment sampling (DUMIS) has been proposed and successfully used for expediting investigation of contaminated sites in the US, Canada and Denmark etc. Yet, it has never been demonstrated in China. The objectives of the present study is to explore the suitability of modified DUMIS in performance evaluation of immobilization of a Ni and Pb contaminated site in China. The demonstration site was a former inorganic salts producer, located in Shandong province, China. Environmental investigation showed that industrial activities and illegal burial of hazardous waste have caused soil contamination of Ni and Pb at the site. Ex-situ immobilization was proposed as the remedy. Contaminated soil was excavated and mixed with immobilizing agent using ALLU sieving and mixing equipment. Treated soil was then piled up for curing followed by confirmation sampling. In compliance with existing Chinese remediation confirmation guidance, decision unit (DU) was set as one DU per 400m² or 500m³ soil. For each DU, a multi increment sample was taken with a hand auger from top 0-20 cm of 50 subsample points evenly distributed within the DU. In one of the field DU, two sets of triplicates were taken, one by DUMIS method (50 subsample points) and the other by the routine method (9 subsample points). To evaluate the field sampling error, triplicates were taken from 10% of DU in the field. The results showed that the relative standard deviation (RSD) of all field triplicates ranged between 3.8% and 66%. The triplicates with high RSD were collected from pit wall or pit bottom with lumps of Ni containing hazardous waste. This study demonstrated that DUMIS method produces high quality data which greatly improved the confidence in remediation performance evaluation at the site.