The bearing capacity of shallow foundations is significantly affected by stratum uncertainty, mainly including geological uncertainty and spatial variability of soil properties. The influence of geological uncertainty and soil spatial variability on the bearing capacity of shallow foundations has been separately investigated in previous studies. This study aims to develop a general probabilistic computational framework to reveal the effects of geological uncertainty and spatial variability of soil properties on the bearing capacity of shallow foundations, in which the geological uncertainty is simulated by Markov random field and the soil spatial variability is characterized using log-normal random field in different strata considering variations of the vertical correlation distance. Based on the borehole and soil data collected from Mawan, Shenzhen, shallow foundation bearing capacity analysis is performed according to the proposed computational framework. The subset simulation method is used to accelerate the calculation of the reliability of each scenario, and reduction factors are proposed to reduce the calculation results to different degrees with the aim of simplifying the consideration of spatial variability. Contribution indexes are defined to quantify the effects of the geological uncertainty and spatial variability of soil properties on the bearing capacity results of shallow foundations. The results show that the traditional deterministic bearing capacity calculation will overestimate the bearing capacity of shallow foundations without considering the stratum uncertainty. When the number of boreholes is sparse, the geological uncertainty has a greater influence on the calculation results; when the number of boreholes is sufficient, it is mainly dominated by the spatial variability of soil properties.