The pipelines are the main transport and distribution system of production in the oil industry, which are subject to environmental and operational conditions which are not favorable for the operation of the pipelines and sometimes represent the risk of accidents and high economic losses. The evolution of risk begins with a systematic search of possible threats to the integrity of the pipeline. The identification of potential threats should not be limited to known risk categories reviewed, but must complete the steps to find new and unique expressions of risk and the study of particular cases. Thus, the importance of a comprehensive risk assessment of the transmission pipeline is crucial. In this paper an integrity assessment for corrosion damages in pipelines was developed through a methodology based on risk analysis to estimate the propagation rate on the time, the size, location and number of damages. To perform this study a data obtained from smart pig runs and an artificial neural network (ANN) model with retro propagation was used. From the data obtained from launching smart pigs on the pipeline it was carried out the training of the neural network, later on it was applied the network previously trained to get the predictions of damages on the pipeline, considering that pipeline did not have any maintenance.