<p>Understanding travel mode choice mechanisms is crucial for achieving balanced and sustainable integrated shared mobility systems. Existing studies have examined a wide spectrum of explanatory factors, including generalized trip cost, socio-demographic characteristics, built environment measures, and psychological or attitudinal constructs. However, existing studies rarely integrate multi-level historical experience measures into interpretable mode choice models, largely because conventional revealed preference (RP) or app-based datasets lack the breadth and consistency needed to construct such indicators. To address this gap, this study extracts generalized cost attributes and experience-related variables from the Baidu Maps navigation dataset, which include user travel frequency, OD flow levels, and their alternative-specific performances. Based on these variables, we develop multinomial logit models (MNL) and optimize model structure for three major transportation modes (metro, taxi, and bicycle-metro) to quantify the effects of various experience-related characteristics on mode choice in integrated shared mobility systems. Our model confirms that users' historical travel experiences significantly influence their mode choices, exhibiting temporal and individual heterogeneity. The frequency of metro use significantly enhances its service stickiness, indicating that incentive policies grounded in historical usage can be vital in fostering and reinforcing user loyalty. In contrast, taxi users exhibit lower mode persistence, likely due to the flexible and less habitual nature of taxi travel. Moreover, time-based pricing policies may effectively prompt temporally sensitive users to adopt the bicycle-metro mode, promoting bicycle-metro as a complementary extension of metro trips and helping ease temporal demand pressure at congested nodes.</p>