This paper presents a PID design method for cascade control systems to tune the parameters of two PID controller directly using the process data collected from one-shot closed-loop test with no need for help from plant models. The design goal is to find PID parameters such that the resulting inner and outer loops behave as closely as possible to the appropriately specified reference models. The optimization problems pertaining to the proposed design are derived. The secondary controller is designed for enhanced disturbance rejection to benefit greatly from cascade control. Without requiring an additional test, the primary controller is designed that accurately accounts for the inner loop dynamics. Two adjustable parameters in the proposed design can be used to deal with the trade-off between performance and robustness for inner and outer loops independently. The desired levels of system robustness explicitly guide the selection of these tuning parameters. Simulation examples show that the proposed method is able to achieve better control performance than the model-based design methods, which confirms the effectiveness of the proposed PID design method for cascade control systems.