Soil pH, as a crucial variable in the soil environment, has a significant impact on the structure and function of terrestrial ecosystems. The Tibetan Plateau is recognized as one of the most sensitive regions to global climate change. However, the degree and direction of change in alpine soil pH and the driving factors remain poorly understood. Here, we utilized resampling methods and machine learning algorithms to explore the spatiotemporal variations and driving factors of soil pH on the Tibetan Plateau from the 1980 s to 2020 s. Our results showed that over the past nearly four decades, changes in soil pH on the Tibetan Plateau have not followed a unidirectional trend of either acidification or alkalization, but instead have been related to the initial pH. There is a significant negative correlation between the change in soil pH and the initial pH value (P < 0.01). Specifically, soil pH > 9.5 exhibited a significant trend of acidification, with an average decrease of 1.55 units, while pH <= 6.5 exhibited a significant trend of alkalinization, with an average increase of 0.32 units. We found that the initial pH was the primary driving factor regulating these changes in soil pH, with a total contribution of 53.13 %, while also mediating the effects of other factors such as topography, climate, and soil characteristics. The initial pH value of the soil may determine its chemical buffering capacity, ion exchange properties, and sensitivity to external acid-base inputs. Our study provides important insights into the mechanisms controlling the spatiotemporal variation of soil pH in alpine regions and highlights the need to place greater emphasis on the intrinsic properties of soil in future soil resource management.