In this paper, we present a learning control algorithm used in our research of advanced opto-electronic automation, which yields high performance, low cost opto-electronic alignment and packaging through the use of intelligent control theory and system-level modeling. The learning loop technique is activated at a lower sampling frequency for specific and appropriate tasks, to improve the knowledge based control model. Our automation technique is based on constructing an a priori knowledge based model, specific to the assembled package's optical power propagation characteristics. From this model, a piece-wise linear inverse model is created and used in the "feed-forward" loop. This model can be updated for increased accuracy through the learning loop.