This paper discusses the use of nonlinear inferential parallel cascade control (NIPCC) to improve control system performance. The goal of NIPCC is to detect and compensate for the effects of unmeasured disturbances faster than the feedback control being employed. NIPCC behaves in a manner that is similar to normal cascade control, but NIPCC has a parallel feedback architecture. NIPCC can be thought of as a limiting case of developing a soft sensor, and thus, NIPCC draws from both parallel cascade control and inferential sensing. The effectiveness of NIPCC is demonstrated on the Tennessee Eastman test process. NIPCC holds promise as an effective means of improving plant-wide control.