BACKGROUND: The internet's low cost and potential for high reach makes Web-based channels prime for delivering evidence-based physical activity (PA) interventions. Despite the well-studied success of internet-based PA interventions in primarily non-Hispanic white populations, evidence on Spanish-speaking Latinas' use of such interventions is lacking. The recent rise in technology use among Latinas in the United States, a population at heightened risk for low PA levels and related conditions, suggests that they may benefit from Web-based PA interventions tailored to their cultural and language preferences. OBJECTIVE: The goal of the research was to examine participant engagement with various features of an internet-based PA intervention for Latinas and explore how use of these features was differentially associated with adoption and maintenance of PA behavior change. METHOD: Pasos Hacia la Salud tested a Spanish-language, culturally adapted, individually tailored, internet-based PA intervention versus a Spanish language, internet-based, Wellness Contact Control condition for underactive Latinas (N=205, mean age 39.2 [SD 10.5] years, 84% Mexican American). These analyses examined engagement with the website and explored how use was associated with adoption and maintenance of moderate to vigorous physical activity (MVPA) behavior. RESULTS: Overall, participants logged on to the website an average of 22 times (SD 28) over 12 months, with intervention participants logging on significantly more than controls (29 vs 14.7, P<.001). On average, participants spent more time on the website at months 1, 4, and 6 compared to all other months, with maximum use at month 4. Both log-ins and time spent on the website were significantly related to intervention success (achieving higher mean minutes of MVPA per week at follow-up: b=.48, SE 0.20, P=.02 for objectively measured MVPA and b=.74, SE 0.34, P=.03 for self-reported MVPA at 12 months, controlling for baseline). Furthermore, those meeting guidelines by the Centers for Disease Control and Prevention for PA at 12 months (≥150 minutes per week of MVPA) logged on significantly more than those not meeting guidelines (35 vs 20 over 12 months, P=.002). Among participants in the intervention arm, goal-setting features, personal PA reports, and PA tips were the most used portions of the website. Higher use of these features was associated with greater success in the program (significantly more minutes of self-reported MVPA at 12 months controlling for baseline). Specifically, one additional use of these features per month over 12 months translated into an additional 34 minutes per week of MVPA (goals feature), 12 minutes per week (PA tips), and 42 minutes per week (PA reports). CONCLUSIONS: These results demonstrate that greater use of a tailored, Web-based PA intervention, particularly certain features on the site, was significantly related to increased PA levels in Latinas. TRIAL REGISTRATION: ClinicalTrials.gov NCT01834287; https://clinicaltrials.gov/ct2/show/NCT01834287.