Objective: To assess the test-retest reliability of a questionnaire to determine nonspecific low back pain (NLBP) and its characteristics in schoolchildren aged 7 to 12 years.Metodology: An evaluation of diagnostic technologies was done in 84 schoolchildren (9.2 ± 1.6 years), 50 % girls, selected for convenience.The questionnaire was administered by individual interview, twice, with an interval of 2 to 4 days.We calculated the Intraclass Correlation Coefficient (CCI 2.1) to quantitative variables and Cohen´s and Weighted Kappa Index to qualitative variables with their respective confidence intervals at 95 % (CI95 %).Results: The prevalence was 20.2 % and 17.9 % in the first and second evaluation, respectively.The reliability of the questions that determine prevalence of NLBP was very good (k = 0.85, 95 % CI 0.69 -1.00, n = 84); The number of days with NLBP in the last month got a fair reliability (CCI = 0.36; 95 % CI -0.009-0.70),while the duration asked by four categories of response was very good (k = 1.00;IC95 % 1.00-1.00),the intensity assessed by Wong-Baker F aces Pain Rating Scale was very good (ICC = 0.82, 95 % CI 0.54 -0.93).In relation to consult a professional for back pain, Kappa indices ranged from k = 0.46 and k = 1.0.In the second evaluation, the children showed shorter duration of pain (3.9 days) and lower intensity (2.0).Discussion: The importance of reliability evaluation of the questionnaire to determine DLI and its characteristics is that there can be a valid measure unless the measure has some degree of reliability.The questionnaire to determine NLBP and their characteristics in schoolchildren aged 7 to 12 years is sufficiently reproducible for use in population studies.