In order to control the weld reinforcement for guaranteeing roller life and increasing rolling quality during continuous cold rolling, the effects of welding parameters on top reinforcement (Rt), bottom reinforcement (Rb), and the thickness ratio of total reinforcement to base metal (Rp) of laser-arc hybrid weld was studied in terms of response surface method. It was found that Rb depends on the laser power and welding speed, while Rt and Rp are determined by welding speed and wire feeding rate. The quadratic regression models of Rt, Rb and Rp were established by the least square method, the fitting degrees of which are 0.928, 0.893 and 0.965, respectively. 5 parameter groups were selected and carried out to verify the models. The experimental data of weld reinforcement well agreed with the predicted results, demonstrating the accuracy of the models. The welding parameters were obtained according to the multi-objective optimization on the basis of welding speed, Rt, Rb, etc., which is significant for parameter selection in industrial applications.