The pretreatment of web data is often the most laborious and requires the most time, this due in particular to the lack of structuration and the large amount of noise present in the raw data. Pretreatment of Web log files is to clean and organize the data contained in these files to prepare them for future analysis. Web log files are often text type, an objective of the pretreatment step is to transfer the data in an easier to use environment (eg in a database). In this paper we will start with the presentation of different formats of web log files, then we will present the different pretreatment methods that we used as cleaning of Web robots queries, removing queries relating to scripts (“.js”, “.css”, ”.swf”) , identifications of users, sessions and visits.