An online resistance grain moisture detector is designed, based on the model of the relationship between measurement frequency and grain moisture and the nonlinear correction method of temperature. The detector consists of lower computer, the core function of which is the sensing of grain resistance values which is based on V/F conversion, and upper computer, the core function of which is the conversion of moisture and frequency and the nonlinear correction of temperature. The performance of the online moisture detector is tested in a self-designed experimental system; the test and analysis results indicate that the precision and stability of the detector can reach the level of the similar products, which can be still improved.