Mapping the universe has always been a salient endeavor in astronomy and astrophysics. Advancements in observational astronomy have generated vast amounts of data containing various features of celestial objects. Inducing a growing need for accurate and detailed classification and localization of stellar objects in the cosmos. In this paper, we present a comprehensive study that combines machine learning techniques to classify celestial objects into distinct categories and predict their precise locations in the sky. This study is divided into two parts: a classification task, where the stellar objects are classified into galaxies, stars, or quasars (quasi-stellar radio sources). The resulting model exhibits exceptional performance in differentiating these objects, as demonstrated by high classification accuracy. We extend our analysis to predict the location of stellar objects using regression techniques. By employing multi-target regression, we model the right ascension and declination coordinates, enabling accurate localization of celestial objects on the celestial sphere. The practical implications of our research lie in producing comprehensive celestial catalogs, facilitating targeted observations, and contributing to the broader field of observational astronomy. The ability to accurately classify and localize stellar objects lays the groundwork for mapping the cosmos and advancing our understanding of the universe's intricate structure.