The precision control of the injection process in micro injection molding demands precise tracking of the moment that melt front passes through the nozzle so that accurate prediction of the amount of melt being injected into the mould can be achieved. The tracking accuracy, to large extent, depends on the processing of pressure signals followed by the identification of the critical moment that melt front reaches the nozzle. A new adaptive Kalman filter was introduced in this study to calculate and predict the amount of melt injected. The filter eliminated the delay error and was more robust than other filters. The adaptive Kalman filter switches between two Qs for steadystate and transient estimations, allowing resetting of the Kalman gain so that convergence is speeded up in calculations. Experimental and simulation results prove the effectiveness of the method proposed.