There is an increasing tendency of industrial sensors becoming intelligent devices when they incorporate additional functions to their primary task of measuring. The embedded systems follow the tendency to use computational intelligence tools in an embedded way in machines, sensors, and actuators, making them intelligent. There are several works in the field of embedded systems, sensors, and pure or hybrid intelligent systems; however, few unite these fields into one single project. This work shows a Takagi Hayashi Fuzzy Neural system for embedding in industrial sensors, making them intelligent devices with the function of automatically detecting calibration errors and self-compensating in operation. The system was tested with data from actual sensors, and the results were analyzed with regard to the speed of execution, frequency of data input in the system, and the number of logic elements used. The results obtained showed that embedding the proposed system in industrial sensors is viable for making intelligent devices. The system proposed was able to self-compensate for errors caused by sensor decalibration and to identify the moment to perform a new calibration after the advanced deterioration of the sensor. The software Quartus II $$^\circledR $$ , from Altera $$^\circledR $$ , was used to develop the project, perform the simulation, and analyze the results.