In recent years, numerous exploration and development instances have conspicuously manifested the tremendous exploration potential of volcanic rock oil and gas reservoirs. The intermediate-basic volcanic rocks in the Changling fault depression exhibit highly distinctive geological characteristics, featuring extensive lateral distribution, complex multi-phase superimposition in the vertical direction, intricate internal structures of volcanic edifices, and extremely strong heterogeneity of reservoirs. Employing accurate low-frequency models to constrain the reservoir phase-controlled inversion has emerged as the key factor restricting the efficient development and utilization of volcanic rock oil and gas reservoirs. In this paper, seismic data is taken as the main variable, the dip-oriented isochronous stratigraphic framework as the first collaborative variable, and the lithofacies indicator curve as the second collaborative variable. The low-frequency model is obtained through interpolation using the multivariate collaborative Kriging method, and the high-resolution volcanic lithofacies body is optimized through waveform indicator simulation. Subsequently, the lithofacies prior control is utilized to construct the joint sampling space of physical properties and elastic parameters, and the physical property prediction of volcanic rock reservoirs is realized based on the Bayesian - Sequential Gaussian Simulation method. This approach fully integrates the data from different dimensions under the existing conditions, effectively reducing the multiplicity of solutions in the prediction results. It supports the deployment of eight well locations in the Longfengshan area, with an average coincidence rate of 81.4% for physical property predictions. This fully substantiates the reliability and validity of this method, laying a solid technical foundation for the subsequent exploration and development work of volcanic rock oil and gas reservoirs.