The invention discloses an adaptive fuzzy inference-based identification method for content of water by using a ultra wide band soil signal, belongs to the technology application direction of fine agriculture, particularly relates to the field of remote sensing technology and detection technology of the content of the water in soil and solves problems that the detection of the content of the water in the soil is inaccurate, the detection cost is high, as an existing method for detecting the content of the water in the soil can not fully use an echo property of the soil in the prior art. The method comprises the following steps: collecting ultra wide band soil echoes of and the content of the water in the soil corresponding to the echoes; preprocessing the soil echoes; building an adaptive fuzzy inference system to carry out eigenvector extraction on the soil echoes after preprocessing; building a classifier by using a machine learning algorithm-random forest algorithm to obtain a random forest classifier; and classifying and identifying extracted eigenvectors on the soil echoes after preprocessing by using the random forest classifier according to the different contents of the water in the soil, and outputting classified and identified results. The method is used for measuring the content of the water in the soil.