The detection of the end of polishing during the chemical mechanical planarization (CMP) process is a critical task in semiconductor manufacturing. Even the research in CMP has grown and developed, one cannot predict the End point detection in order to avoid over or underpolishing wafer. The disadvantages of offline approach in endpoint detection have incited the researchers to discover an efficient substitute. In this Master thesis, an alternative approach named online method has been presented in which a sequential probability ratio test (SPRT) was developed and applied to the wavelet decomposed Acoustic emission data collected during the progression of the CMP process. Two dispersion parameters were used to detect endpoint, the first one standard deviation and the second the coefficient of variation. This test is shown to be efficient in controlling complex processes and appropriated for real-time application by developing a moving block strategy. This book is useful in industrial statistic especially in process control and also for in the field of the semiconductor production especially in CMP process.