<p>The application of geospatial big data to analyze temporal variations in tourist flows and their relationship with fine-grained mobile search engine data merits further exploration. This study explores intraday and inter-day patterns of tourist flows and their correlations with pre-trip online search activities by integrating mobile signaling data and long-term pre-trip search data from multiple platforms. Focusing on Zhongshan Mountain over a three-year period, the research employs the DCC-GARCH method to analyze nonlinear dynamic intraday correlations stemming from different clients (mobile vs. PC), and the TVP-VAR model to examine time-varying impact mechanisms. Key findings reveal: (1) Tourist flow and search volume sequences exhibit a peak-thick-tail distribution with ARCH effects and weekly fluctuations. (2) Pre-trip search fluctuations across devices correlate with tourist volume changes, with mobile-terminal correlations being high and stable, while PC-terminal correlations show positive associations two days in advance, varying seasonally. (3) Weekly fluctuations influence precursor effects, with search indices from 1 to 8 days prior significantly predicting tourist flows, particularly strengthening over 1–4 days. A 1% increase in the mobile search index results in a 0.1%-8% rise in tourist volume across lag periods. By integrating geospatial big data and temporal analysis, this study reveals shifts in pre-travel search behaviors and the spatio-temporal compression effects driven by new technologies in the mobile Internet era. By correlating tourists’ online search activities with their physical movements, the research provides strategies for technology-enabled tourism marketing and destination capacity optimization.</p>