This work addresses the adoption of artificial neural networks to pre- dict the monthly precipitation volume in the city of Manaus. For that, we consi- dered data from 1950 to 2015 and climatic indexes of sea surface temperature that influence precipitation in the region of interest. Different neural networks were proposed, trained and tested for the problem considered and, for each month of the year, the network with the best accuracy was selected. The results obtained show the adequacy of the proposed models, with accuracy higher than 92%, emphasizing the advantages in using Machine Learning in this scenario.