Abstract: Tight sandstone reservoirs exhibit low porosity, low permeability, and more complicated relationships between seismic elastic parameters and reservoir physical parameters than conventional reservoirs, which result in challenging rock physics deterministic modeling and inversion.Based on the analysis of reservoir physical properties, a prediction process for porosity, pore scale, and permeability of tight sandstone reservoirs was built, which used the kernel-Bayes discriminant method.First, considering the relationship between the permeability and the pore scale, an equivalent pore scale method was proposed.Then, an analysis of reservoir physical properties was performed to identify the sensitive elastic parameters.Next, the kernel-Bayes discriminant method was adopted to predict the petrophysical properties; logging and seismic data were used to verify the feasibility of the method.The predicted porosity and permeability were well-matched with the well-log data.Thus, we determined that this method could effectively identify the sandstone reservoir and describe its porosity and permeability.The proposed prediction method for physical properties in tight sandstone reservoirs is significant for the exploration and development of oil and gas fields.