Modeling the spatial distribution of reservoir geological and petrophysical properties plays a vital role in characterizing and forecasting its performance. Building meaningful reservoir models is an outcome of the integration of various sources of information such as geological, engineering, and seismic data in the modeling process in an optimal manner. In this study, a new method is proposed for generating and updating the reservoir lithofacies models using well and seismic data. The proposed method is a combination of the Particle Swarm Optimization (PSO) and the Sequential Indicator Simulation (SIS) geostatistical technique. Quantitative and qualitative analysis of the obtained results on a 3D synthetic model demonstrated that the use of the proposed method led to the construction of the reservoir lithofacies models with an increase of 13% (on average) in accordance with the reference lithofacies model compared to the conventional geostatistical approaches.