The demand for more economic and efficient processes may lead the plant to operate at several conditions. In this context, the use of automatic control structures for successful and safety operation is paramount. The control system performance depends on the quality of the processes data. However, the sensor devices and measurement procedures are important uncertainty sources. For effective control actions, one alternative considers reducing the measurement noise so as to keep the operation at desired conditions. In this regard, this work studies the on-line use of particle filters to improve the input data of the controllers. The proposed scheme should estimate the actual states from noisy measures before determining the control action. A feedback temperature control was taken as an example. The results show that the scheme is a potential tool for on-line application in real process systems, since the computation time taken by the filter is consistent with usual sample times.