The micromobility mode of transportation is convenient for short-distance trips in urban settings. The popularity of dockless sharing systems is rapidly increasing because of their characteristics of convenient rental/return and mobile payment options. Although many previous studies have examined travel behaviors regarding a single means of micromobility transportation, few have analyzed mode choice behavior among multiple, similar, and competing means of micromobility transportation. To our knowledge, no researcher thus far has examined micromobility as a potential means to solve the last-mile problem from the subway station to the destination. Therefore, a stated preference survey was conducted, analyzing factors and preferences influencing the mode choice of subway users for the last-mile trip. This study considered sharing dockless e-scooters and docked bicycles, as well as walking. It analyzed the differences in mode choice behavior of subway users based on their demographic and socio-economic attributes, as well as travel (travel purpose and distance), mode-specific (fare), and environmental (slope) characteristics. Two logit-based models—the multinomial logit (MNL) and mixed logit (ML) models—and two ML models based on residential areas of Seoul and Gyeonggi, Korea, were developed and compared. The results showed that the ML model was superior to the MNL model regarding model fit and individual parameters. In the ML model, the statistically significant variables included fare, marital status, household size, age, income, disability, and slope of roadway. However, the significance differed by alternative modes: for e-scooters (vs. bicycles), the factors influencing last-mile mode choice behaviors showed greater statistical significance. A comparison of the two ML models by region highlighted major differences in the results for the variables of household size and trip purpose between residents of Seoul and Gyeonggi. Urban transport policymakers interested in introducing micromobility services can use the findings of this study to make more well-informed decisions about traffic congestion problems and increase public transportation revenue in an environmentally friendly manner.