Epidemiological patterns of malaria are influenced by different kinds of climate. In Sahelian countries, the link between climatic factors and malaria is still insufficiently quantified. The aim of this work was to conduct a time-series study of rainfall to estimate the increased risk of malaria morbidity. Daily suspected malaria episodes among subjects of all ages were collected retrospectively in three health care facilities between 1 January 2000 and 31 December 2003 at Magaria, Niger. These daily numbers were analysed with time-series methods, using generalized additive models with a negative binomial family. The impact of rainfall 40 days before occurrence of suspected malaria episodes was studied using a distributed lag model. More than 13 000 suspected malaria episodes were registered corresponding to an annual cumulative incidence rate of 7.4%. The overall excess risk of suspected malaria episodes for an increase of 1mm of rainfall after 40 days of exposure was estimated at 7.2%. This study allowed to specify the excess risk of rainfall on the occurrence of suspected malaria episodes in an intermediate rainfall area located in the Sahelian region in Niger. It was a first step to a health impact assessment.