An adaptive learning algorithm using laser velocity meter is proposed for the process control of tandem cold mills. This method is used for on-line adaptive learning calculation of deformation resistance of strip and rolling friction models. The reverse calculation values of deformation resistance and friction coefficient as variables of non-linear equations can be obtained by introducing the measured rolling force and forward slip into the calculation model individually. Using Newton-Raphson iterative calculation to solve the non-linear equations, the reverse calculation values can be calculated. The actual forward slip obtained from the measured strip rolling speed using laser meter makes it possible that the deformation resistance of strip and friction coefficient can be uncoupled for model adaptive learning calculation. The inverse calculation and adaptive learning can improve whole control system of tandem cold mills significantly.