Filtration is a procedure for changing raw water into clean water using filtering media. Over time, the quality of the filtration media will decrease, therefore maintenance needs to be carried out. This research develops a decision support system for membrane filter maintenance in portable South Kalimantan peat water treatment solar-powered, using the Fuzzy Inference System (FIS) Takagi-Sugeno-Kang method, which is implemented in a microcontroller system using three input variables, namely turbidity, TDS, and filtration pressure. Based on the tests that have been carried out, the microcontroller system design has provided information on the filter's condition, and the pump actuator can work automatically to determine the appropriate condition of the filter based on changes in the set variable values. The sensors used were tested to determine the accuracy of their readings. Testing the system design compared with MATLAB calculations obtained an accuracy of 96.47%.