Intense and long-lasting rains often cause troubles to the population in many cities and to proper urban planning and civil defense activities it is necessary to know the expected precipi1 Biometria e Estatística Aplicada, Univ. Federal Rural de Pernambuco-UFRPE, cavalcanti.rodrigo@gmail.com 2 Biometria e Estatística Aplicada, Univ. Federal Rural de Pernambuco-UFRPE, chsestatistica@gmail.com 3 Biometria e Estatística Aplicada, Univ. Federal Rural de Pernambuco-UFRPE, elielmas93@gmail.com 4 Doutor em Agronomia, Univ. Federal Rural de Pernambuco-UFRPE, moacyr2006@gmail.com 5 Doutora em Física, Univ. Federal Rural de Pernambuco-UFRPE, tastosic@gmail.com tation. In this work, we pursued to estimate the probable maximum precipitation (PMP) daily and accumulated from 2 to 7 days for Recife-PE and to estimate the return time for the PMPs found. Moreover, based on the theoretical distribution function of the data, calculate the PMPs for 2, 5, 10, 20, 50, 100 and 200 years. The pluviometric data were captured by the data collection platform (DCP) of the National Institute of Space Research (NISR) for the period 1998 to 2012. The adjustment of the Gumbel distribution was observed through the Kolmogorov-Smirnov test at the level of significance of 5%, the parameters of this distribution were obtained by the moments method. The estimates for the PMPs were determined by the Hersfield method, where it considers the frequency of annual maximum rainfall data recorded at the site of interest. Excel and R software were used for the analysis. The results showed that in the accumulated 2 days, the return time of the PMP of 548mm is 82 years. It was estimated that every 2 years, rainfall events occurring at or above 125mm in 24 hours and that every 5 years rainfall events occurring at or above 211mm in 24 hours. Such results should support policies of civil defense and urban planning to mitigate damages to mobility, habitation, economy and life.