The method discloses a method of using seismic data to carry out mud shale gas content prediction. The method comprises the following steps of (1) measuring data, wherein geological data includes well logging, gas testing and mud shale core gas content testing data; logging data includes interval transit time, a density and a gas content logging interpretation curve passing through a core gas content testing data scale; and seismic data includes seismic materials of conventional post-stack or prestack offset processing; (2) using high correlation of a density logging curve and mud shale gas content data to establish a gas content calculating model based on the density logging curve, wherein the model satisfies the following equation: VGAS=f (DEN) (1); the VGAS is a gas content, m/t; the DEN is the density logging curve, g/cm; and f is a function equation between the gas content and the density; (3) carrying out post-stack multi-attribute inversion and acquiring a density curve data volume required by the gas content calculating model; and (4) carrying out gas content prediction: substituting the density curve data volume acquired from the step (3) into the equation (1) to acquire a mud shale gas content.