Big data has become a primary resource for decision-makers. Data analysis makes it possible to extract new data which can be very useful. Meanwhile, the data owner has to protect the privacy of people and establishments. Therefore, private information can neither be all shared nor extractable from any analysis. That is why the owner of the data must negotiate with the explorer according to the objectives of each of the two. In many cases objectives can contradict each other. The challenge is to make a compromise between protecting privacy and retaining the usefulness of the data processed. This article strives to propose a model of anonymization based on the theory of fuzzy logic. In the process of anonymization, we assign, in the first step, a degree of identification and sensitivity to each data, then the qualitative data is encoded before we finally anonymize depending on the type of data. Our model has been proved to be useful and reliable for both the owner and the explorer of the data. Therefore, it helps protect the data and ensures its privacy.