Purpose: We have chosen the pupils of Grade two to four from ten primary schools in Shanghai to target high-risk children of ADHD for the further multi-assessment. Tools: SDQ (parent-report) was completed by the American psychologist Goodman. R. and the norm in Shanghai was completed by Kou Jianhua, Du Yasong and Xia Liming in 2005, which shows good reliability and validity. Cross-cultural applicability of SDQ proved to be good. Methods: The sample consisted of 4483 children (7–10 years old) with valid answers on the whole 25 items of the SDQ parent-report. IBM SPSS Statistics 19 was used for statistical treatment. According to the results of statistical treatment, the number of the final selected subjects was 4454, aged from 7 to 10, including 2297 boys and 2157 girls. These subjects were divided into eight groups by age and sex. Boxplot, Descriptive Statistic and Independent-Samples T Test were used to for data processing. After that, we classified the scores of ADHD by groups (see Table 4 and 5). Procedure：In September of 2009, we informed the parents of the objects and obtained their consent of SDQ. Then 4483 valid questionnaires were received by deleting 305 invalid questionnaires from the total 4788. Effectiveness of the questionnaires is up to 93.63 percent. Results: Among boys, 8-year-old subjects achieved the highest score and 9-year-old subjects showed the most decentralized degree of the distribution; among girls, 9-year-old subjects achieved the highest score and showed the most decentralized degree of the distribution (see Table 1). Boys had higher scores (p<.01) at each age (see Table 1 and 2). Differences in ages existed but weren’t significant (see Table 3). Girls and 9-year-old boys eventually showed five grades, for there were no ADHD scores below the mean by two standard deviations. The remaining subjects receive six grades (see Table 4). Besides, in Grade Red and Orange, the two highest, the total percentage of girls was higher than boys (4.50VS.2.92, 12.48VS.11.48). Conclusion： (1)The most serious symptoms of ADHD appear at 9; (2) Girls have a higher risk of ADHD; (3) Children with High-risk ADHD are screened out: boy 14.4%, girl 17.4%. Innovation：The usage of hierarchical Strategy proved to be more scientific, with the research data reflecting the objective situation of ADHD with more details and the research findings closer to the truth.