To date, more than 150 exo-solar planets have been observed by various methods such as spectroscopic, photometric, astrometric, gravitational lensing, pulsar timing methods. However, all these are indirect methods; they do not directly image the planets. Only free-floating planets or their 'ana-log' have been directly detected so far. Thus the next milestone is the direct imaging of any kinds of planetary mass objects orbiting around normal (young) stars, which might have been associated with protoplanetary disks, the sites of planet formation. I will describe some SUBARU efforts to detect self-luminous young giant planets as companions as well as direct imaging of the protoplanetary disks of ${\~}$ 수식 이미지100 AU size. The results of near-infrared coronagraphic imaging with adaptive optics are briefly presented on AB Aur, HD 142527, T Tau, and DH Tau. Our results demonstrate the importance of high-resolution (${\~}$ 수식 이미지0.1 arcsec) direct imaging over indirect observations such as modeling based on spectral energy distributions. The SUBARU observations are a prelude to ALMA from the morphological point of view.