The statistical process monitoring is usually carried out through two phases: Phase I and Phase II. These phases are different in term of aims and performance metrics. When the quality characteristic of a particular process or product can be described by a linear regression model between response and independent explanatory variables, we are dealing with a “profile” which is categorized in terms of the structure of regression model. The multivariate methods are usually applied for profile monitoring. In most of the proposed Phase II monitoring methods, the fundamental assumption is that the process/profile parameters are known. While this assumption is not usually valid in practical situations and the process/profile parameters are estimated based on in-control samples collected in Phase I. This study aims on evaluating the effect of parameter estimation in Phase I on the performance of Phase II control charts for monitoring multivariate simple linear (MSL) and multivariate multiple linear (MML) profiles in terms of Average of average run length (AARL), standard deviation of ARL (SDARL) and coefficient of variation of ARL (CVARL). We worked on three well-known Phase II control charts for monitoring MSL profiles and also four Phase II control charts for monitoring MML profiles. According to the previous studies on parameters estimation effect, we expect that the parameters estimation strongly affects the Phase II performance. One of the goals of this research is determining the monitoring methods which is less affected by Phase I estimation. It is not surprising that superior methods based on estimated parameters would not be the same as that of based on known parameters. Moreover, a new approach for determining the optimal number of Phase I samples is proposed so that yields an estimation with proper estimation in terms of Phase II performance metrics. For this aim, we proposed two optimization models based on Phase II metrics and solved them by simulation-based optimization approach. In addition, a new cluster-based approach is developed in order to improve Phase I estimation. The performance of the newly proposed approach on Phase II performance is compared to the competing approach in terms of AARL and SDARL. The Bayesian theory has been recently applied for profile monitoring in a few studies. According to this theory, the prior information can be incorporated for estimating the model parameters. The results of previous studies denote very good potential of Bayesian estimator for improving the Phase II performance of control charts. Therefore, in the final chapter of this research, we developed one Bayesian multivariate exponentially weighted moving average (BMEWMA) control chart and two Bayesian multivariate cumulative sum (BMCUSUM) control charts for monitoring multivariate linear profiles. The simulation results illustrated the superiority of Bayesian control charts in detecting shifts in regression coefficients rather than the ltr"