The shearing machine is an important and complex accessory equipment of the continue-mode rolling mills. Its mechanism design scheme determines the shearing quality of steel. The shearing machine mechanism design (SMMD) contains multi conflicting technical requirements and belongs to a multi objective optimization problem with the nonlinear constraints. Recently, ant colony optimization (ACO), a swarm based computing methods, has demonstrated its superiority in many complex optimization problems. This paper presented a quasi TSP-based SMMD model and an ACO algorithm for the SMMDP. The presented method dispersed the searching space of the design variables by setting several different search steps, and an ACO algorithm was adopted to search the best searching step of each design variable dynamically during the whole optimization process. Computational results showed that the proposed method can improve the computational accuracy and produce better solutions within short running times.