Recent astrophysical advancements highlight the challenges in exploring exoplanets, primarily detected through starlight variations. While the Kepler space telescope aimed to identify Earth-like exoplanets around sun-like stars, manual analysis of light curves remains time-intensive. Thus, efficient detection methods are crucial. Leveraging machine learning and deep neural networks, particularly artificial neural networks (ANNs), shows promise for exoplanet detection by discerning intricate data patterns, notably in light curves. Integration of the NASA Exoplanet Archive Dataset with these techniques allows for assessing exoplanet habitability. This approach enhances researchers’ abilities to detect and characterize exoplanets, broadening the quest for habitable worlds beyond our solar system. The study’s innovation lies in merging machine learning and deep neural networks with extensive datasets, particularly the NASA Exoplanet Archive Dataset, to improve exoplanet detection and characterization. It investigates the potential of ANNs in identifying exoplanets and predicting their habitability, offering unprecedented insights into discovering habitable planets beyond our solar system. Notably, using ANNs achieved 88.3% accuracy in exoplanet detection on test data, while gradient boosting yielded 91.06% accuracy in habitability prediction.