In traditional quality control, the quality of product is typically modeled as the univariate or multivariate distribution of the quality parameters. In more recent applications, the quality is modeled using the relationship between a response and independent variable. This paper investigates the performance of multivariate version of the double dMEWMA statistics in detecting changes in step shift in the intercept, slope, and error-variance of simple linear quality profiles. The statistical performance of the dMEWMA chart is estimated and compared versus three different charting techniques includeing the Hotelling T² , EWMA/R and dEWMA3. For all the compared charts, the average run length (ARL) under wide range of shift levels are estimated in order to draw beneficial conclusions.