O presente trabalho teve como objetivo (i) avaliar o processo de controle da qualidade relacionado às vacinas contra febre amarela utilizadas pelo Programa Nacional de Imunizações (PNI) do Ministério da Saúde no período de 2000 a 2008, através de levantamento de dados provenientes do Sistema de Gerenciamento de Amostras (SGA) do Instituto Nacional de Controle de Qualidade em Saúde (INCQS) da Fundação Oswaldo Cruz (Fiocruz) e (ii) propor a utilização de gráficos de controle, como ferramentas úteis para a melhoria desse processo, a partir da análise comparativa dos resultados obtidos para vacinas nacionais e importadas, e da consistência de produção e detecção de tendências sistemáticas. Teve ainda como proposta (iii) a validação retrospectiva do teste de potência da vacina em questão, através da demonstração de sua confiabilidade (variabilidade e precisão) e do acompanhamento do desempenho da vacina de referência usada nos testes de potência e termoestabilidade. Compuseram a análise relativa ao controle da qualidade das vacinas contra febre amarela os seguintes parâmetros: (i.) análise do protocolo resumido de produção e controle emitido pelo fabricante, (ii.) teste de potência, (iii.) teste de termoestabilidade, (iv.) teste de determinação de ovalbumina, (v.) teste de esterilidade bacteriana e fúngica, (vi.) teor de umidade residual e (vii.) endotoxina bacteriana. Para confecção dos gráficos de controle foram estabelecidos os limites de controle (3 DP a partir da média) e limites de alerta (2 DP), calculados a partir dos 20 primeiros resultados a cada ano, como definido pela OMS (OMS, 1997) e através do programa SPC Explorer RT, versão 5.21 (Quality America Inc.). No período estudado ingressaram no INCQS 1031 lotes de vacinas contra febre amarela, produzidos por dois fabricantes distintos (97% pelo fabricante A e 3% pelo fabricante B), representando um total de 285 milhões de doses individuais.The present study aims at (i) evaluating the quality control process - from 2000 until 2008 – of the yellow fever vaccines used by the Brazilian Ministery of Health in the National Immunization Program (PNI), with data generated from the Sample Management System (SGA) of the National Institute of Quality Control in Health – INCQS - Oswaldo Cruz Foundation and (ii) proposing the use of control charts as useful tools to improve the process, using comparative analysis of the results obtained with Brazilian and imported vaccines, production consistency and sistematic tendencies. The other purpose (iii) is the retrospective validation of yellow fever vaccine potency test, with the demonstration of its reliability (variability and precision) and the following up of the reference vaccine performance used in the potency and thermostability tests. The following parameters were used in the analysis of yellow fever vaccines: (i.) the manufacturer’s resumed production and control protocol, (ii.) potency assay, (iii.) thermostability assay, (iv.) ovalbumine determination assay, (v.) bacterial and fungal sterility assay, (vi.) residual moisture content and (vii.) bacterial endotoxin. For the control charts, control limits were established from the calculation of 3 DP and for the alert limits, from 2 DP, after the first 20 results of each year, using the SPC Explorer RT, version 5.21 (Quality America Inc.), as determined by WHO (WHO, 1997). From 2000 until 2008, INCQS received 1031 yellow fever vaccine lots, produced by two diferent manufacturers (97% manufacturer A and 3% manufacturer B), representing a total of 285 million doses. Five lots produced by manufacturer A (0.5%), representing 1,7 million doses - were considered unsatisfactory: 3 due to bacterial and fungal contamination and 2 due to inadequate residual moisture content. Evaluation of the potency and thermostability control charts demonstrate that these assays present less variation at INCQS than at the manufacturer A. In 2007, a significant number of results were bellow the inferior limit of control. These values were concentrated in the second half of the graphic, when, coincidentally, there was a great demand of the vaccine, due to the increase of the number of cases of yellow fever in Brazil and in other South American countries. At INCQS, the process is under statistical control and the special causes of variation, in case they occur, are under control. Regarding the retrospective potency assay validation of yellow fever vaccines, we concluded that the accuracy was 0,19 (0,09-0,25) and recovery was 96,14%. The total precision was 14,48%; repetitivity was 11,92%; and intermediate precision 7,78% between assays and 7,8% between analists. These values were satisfactory for biological assays according to international agencies.