This communication examines three strategies of predictive lubricant monitoring and replacement, used for farm tractors or similar vehicles. These strategies optimise the draining periodicity. They are the off-line follow-up, the sensors follow-up and the analytical model follow-up. The implementation of the suggested analytical model will be discussed, on the basis of field collected data (on a series of tractors, either customer's or on loan). Regular oil samples, and significant ones carried out at the end of the study, were taken and analysed in order to predict the evolution of the lubricant characteristics. Extensions to the experimental study were carried out at the end of this work. They are discussed in the paper (FZG gear scuffing, 4 ball wear and EP…).