Mapping and monitoring invasive plant species are crucial for protecting native biodiversity and enhancing effective conservation efforts. A biologically invasive tree, Acacia auriculiformis C. Cunn. ex Benth., is widespread throughout one of the largest national parks (50.2 km²) in Bangladesh. The species is evergreen, and contrasts with the dominant native deciduous vegetation. Therefore, we examined the phenological information derived from multi-spectral remotely sensed data and extracted phenological spectral features for mapping invasive species distribution. We developed temporal profiles of three vegetation indices (VI)- Normalized Difference VI (NDVI), Green Red VI (GRVI), and Atmospherically Resistant VI (ARVI) to analyze the spectral response of invasive Acacia auriculiformis, dominant Shorea robusta and mixed vegetation using Sentinel-2A (10 m resolution) images acquired from 2022-2023. Based on these profiles, two distinct phenological periods- green and senescence were identified. The VIs from each identified phenological period were processed by using median composite method in Google Earth Engine (GEE). We computed phenological vegetation indices from the composited VIs. These indices were then used to generate a new feature image that contains spectral characteristics from both the green and senescence periods. The approach offered for comprehensive spectral separation of A. auriculiformis from native vegetation. We integrated the feature image generated by phenological vegetation indices into three classification algorithms: Random Forest (RF), Support Vector Machine (SVM), and Classification and Regression Tree (CART) and compared its effectiveness in classification. The classification models were trained and tested using reference data collected from field surveys in 2022. Among the classifiers, RF achieved the highest overall accuracy (87%) with a kappa coefficient of 0.83. The results demonstrated that the composited image containing phenological spectral features significantly improved the spectral separability of the invasive species from coexisting vegetation. By capturing variations in spectral reflectance, this approach enhances classification accuracy. The findings underscore the significance of integrating phenological spectral features to improve forest conservation efforts and achieve accurate mapping of A. auriculiformis.