One of the critical stages in a display production line is the image alignment of displays that includes the precise adjustment of the geometric parameters and the color of the image. The mutual influences of the parameters of the display's image necessitate a complex and interactive alignment process. In this paper, the effect of the mutual influences of geometric parameters on the alignment process of a display's image are shown, a suitable model for the geometric characteristics of a display's image are suggested and, then, the unknown parameters of the proposed model are estimated by the RLS estimator. Using an off-line estimator, an initial measure of the values of the unknown parameters of the display model is obtained. To modify the model parameters of the consecutive video displays on the production line, an on-line estimator is applied. Variations of the parameters of the display model on a production line are traced using on-line estimation. This model estimation is used to implement an adaptive alignment algorithm. Both the adaptive and the proportional alignment processes have been experimentally implemented under similar working conditions. Experimental results show that the use of the adaptive alignment process considerably increases the speed and reliability of convergence of geometric parameters to their desired values. An IA-32, 3.4 GHz Pentium P4 processor has been used in this research. Considering the rapid developments in UDSM technology and the IA-64 architecture, the application of the proposed adaptive alignment algorithm in an auto-alignment system has the potential for real-time implementation in the near future.