There are growing demands for Machinery health monitoring and fault diagnosis of rotating machinery to lower unscheduled breakdown. Gearboxes are one of the fundamental components of rotating machinery; their faults identification and classification always draw a lot of attention. However, non-stationary vibration signals and low energy of weak faults makes this task challenging in many cases. Thus, a new fault diagnosis method which combines the Empirical mode decomposition (EMD), time frequency method, and self-organizing feature extraction machine learning approach is proposed in this paper. In particular, efficient feature extraction and feature selection is a key issue to automatic condition monitoring and fault diagnosis processes. To focus on such issues, this paper presents a research study to formulate an real-time prediction method using vibration signals of various gears conditions that uses empirical mode decomposition (EMD) technique to extract features and select the predominant features by using various classification algorithms.