Fitness apps are effective to enhance health behaviour. However, high dropout rates are observed. Facing disclosure of highly private data, trust in fitness apps can represent a central variable in this context predicting fitness app usage. Trust has been defined as dependence under risky conditions and has been adapted to traditional computer programs The model comprises propensity to trust, institution-based trust, and knowledge-based trust (functionality, reliability, helpfulness). It was the aim of this study to test and validate trust in new technologies by the example of fitness apps. Using a German version of the Trust in a Specific Technology questionnaire, propensity to trust and institution-based trust were analysed in fitness app users (n = 248) vs. non-users (n = 338), and knowledge-based trust was analysed in users vs. dropouts (n = 168) via an online survey. Structural equation models (SEM) and measurements of invariance across the different samples were conducted. In SEM, analyses indicated good model fit (CFI = .95â€“.99) and invariance across groups. Propensity to trust was higher in users compared to non-users (3.54 < z < 3.40, p < .001), whereas institution-based trust was not. All scales of trusting beliefs were higher in users compared to dropout (3.48 < z < 8.25, p < .001). Additionally, low levels of functionality beliefs led to lower duration of fitness app usage (? = .61, p < .001). This study provided evidence for good construct and predictive validity regarding trusting beliefs in fitness apps, but limitations in validity of institution-based trust.