This paper presents an extended evaluation of a flight trajectory prediction framework that integrates Automatic Dependent Surveillance-Broadcast (ADS-B) data with surface wind information. In our previous work, we focused on limited temporal ranges or route-specific datasets. This work investigates the proposed prediction approach using six independently constructed datasets covering different observation periods and data characteristics. Wind direction and wind speed are incorporated as auxiliary explanatory variables to examine their influence on trajectory estimation accuracy. A random forest regression model is used and prediction performance is analyzed using the Mean Squared Error (MSE). The experimental results show that the contribution of meteorological factors varies depending on dataset composition and temporal coverage, highlighting the importance of dataset diversity in evaluating model robustness. This study provides additional insights into ADS-B based flight prediction under heterogeneous operational conditions.