In industry, many nonlinear processes can be approximated well by a linear model under a suitable design parameter, upon which a linear controller can effectively control these processes. The key problem is how to find this suitable design parameter, which is never considered in process design. In this paper, a novel design for control approach is proposed to design the process to have a satisfactory linear approximation. First, a subspace-modeling-based nonlinear measurement is proposed to avoid the infinite dimensional optimization problem in the traditional nonlinearity measurement. Then a particle-swarm-optimization-based design approach is developed to obtain the optimal design parameter through solving the nonconvex and nondifferential measurement problem. Finally, the proposed design is applied to design a practical curing oven and compared with the existing design.