Bruxism is the excessive grinding of the teeth or excessive clenching of the jaw. Early diagnosis of Bruxism is advantageous, due to the possible damage that may be incurred and the detrimental effect on quality of life. A diagnosis of Bruxism is usually made clinically and is mainly based on the person's history e.g. reports of grinding noises and the presence of typical signs and symptoms, including tooth mobility, tooth wear, indentations on the tongue and pain in the muscles of mastication, The neuronal activity of brain Electroencephalogram (EEG) is a highly non stationary signal. For analysis purpose it is useful to divide the EEG into segments in which the signals can be considered stationary. Hilbert Huang Transform(HHT) is an effective tool to understand the nonlinearity of the medium and nonstationarity of the EEG signals. The signals in the frontal plane from electrodes F4C4, FP2F4, F8T4, FP1F3, F3C3 and F7T3 are used to understand and diagnose Bruxism. In this paper Empirical Mode Decomposition (EMD) is used to decompose the EEG signal in to Intrinsic mode functions(IMF). Since some nonlinearity still exists in the intrinsic mode functions, we used non linear analysis methods of IMF's to predict the Bruxism. Largest Lyapunov exponent, Hurst component and correlation dimension of each intrinsic mode function are found. The mean amplitude of the instantaneous frequency of each IMF is also used in the analysis of the signal and the results used in diagnosing the presence of Bruxism.