Adaptive soft sensors in building inferential control systems is a doctoral research that was conducted at the university of York - UK. The research introduces the development of data-driven models based on Gaussian processes, and experimentally demonstrates the integration of these models with control systems to build what is called inferential control. It investigates Gaussian processes as a soft sensor building technique, addresses the issues associated with this method, proposes deploying these sensors in inferential control, and adopts a methodology to enhance the reliability of the introduced inferential control system.