The main objective of this work is to estimate the epoch or GCI (Glottal Closure Instant) locations from the speech signal using a decomposition method, DMD (Dynamic Mode Decomposition) which assumes the speech production system as non-linear. This work also contributes in estimating instantaneous pitch frequencies of the modes obtained from DMD. DMD algorithm captures the center frequency close to the fundamental frequency defined for each glottal cycle of speech utterances through its modes. Based on the obtained mode frequency, a sine wave will be constructed and negative-to-positive zero-crossings will be estimated as epoch locations. Experimental result shows that the DMD approach is very effective in identifying the epoch locations more accurately when comparing with the existing methods.