A horizontal-stress vector prediction method, which relates to the technical field of
in situ stress prediction. The method comprises: calculating elastic parameters and anisotropic parameters of a target reservoir; using a
horizon interpretation for the target reservoir as a constraint, and using the calculated elastic parameters and anisotropic parameters to calculate
logging curves of Young's modulus, Poisson's ratio and the fracture density; on the basis of the calculated
logging curves of Young's modulus, Poisson's ratio and the fracture density, constructing an initial azimuthal elastic impedance model; on the basis of the initial azimuthal elastic impedance model, inverting Young's modulus, Poisson's ratio and the fracture density; substituting into a prediction formula the fracture density and Poisson's ratio, which are obtained from the inversion, and the vertical
principal stress of the target reservoir, and performing calculation to obtain the maximum and minimum horizontal stresses of the target reservoir; and on the basis of least-squares
ellipse fitting, obtaining the directional distribution of the maximum and minimum horizontal stresses within the target reservoir, so as to realize quick, accurate and highly operable horizontal-stress vector prediction. The present invention further relates to a horizontal-stress vector prediction apparatus, a
computer device, and a
machine-readable storage medium.