This paper presents an experimental investigation of the reliability and accuracy of determining the natural frequency of a laboratory-scale bridge model using a smartphone-based drive-by monitoring approach. The study was conducted through two sets of experiments designed to evaluate the influence of various measurement configurations on the accuracy of frequency identification. In the first measurement series, the effect of vehicle speed and the smartphone’s position relative to the vehicle axles was examined. The smartphone was alternately mounted above the first and second axles of a specially adapted test vehicle to assess how its location influences the identified bridge natural frequency under different driving velocities. Based on the obtained results, the configuration that provided the most accurate frequency estimation was selected for further testing. The second measurement series focused on analysing how the vehicle’s lateral position on the bridge deck affects the precision of natural frequency identification. Additional experiments were performed with two vehicles moving simultaneously across the bridge to observe the combined effect of multiple loads. Acceleration data collected via the smartphone were processed in MATLAB, and the identified dynamic parameters were compared across all test scenarios. The results confirmed that both vehicle speed and sensor position influence the accuracy of the extracted natural frequencies. These findings demonstrate that smartphones can be effectively used for low-cost, non-invasive dynamic testing of bridges, if measurement conditions are carefully controlled. The study contributes to improving the practical understanding of drive-by monitoring approaches and supports their further development for reliable, large-scale Structural Health Monitoring (SHM) applications.