Real-time monitoring of a drilling process is an essential task in improving their performances. Faults that might occur have to be detected as soon as possible in order to preserve drilling efficiency. In this paper, drilling process monitoring by identifying time varying parameters through marginalized particle filter (MPF) is treated. The idea consists in enhancing the tracking ability of parameters change by integrating into the process model a part that represents the faulty process and another when the process is safe. The efficiency of the developed approach is highlighted through simulated and experimental data obtained from tests campaign.