In the process of measuring the sediment concentration in flow-water, the temperature will greatly influence the testing results. A new method, based on artificial neural network under MATLAB environment is designed to fuse the two-sensor information. The neural network used the capacitance sensor and the temperature sensor output as input. After the network is trained, it can make the sensor eliminate the temperature influence and improve the target parameter input. The simulation results show that the fusion results are stable and precise enough with this method.