In conventional far-field optical microscopy, it has a limitation on spatial resolution due to diffraction barrier, known as Abbe’s diffraction limit. In detail, when light interacts with an object, interacted light generates near-field and far-field components. The far-field component is propagating light through space and used in conventional optical microscopy. The near-field component is nonpropagating light whose amplitude exponentially decays as it propagates. Because the near-field light includes high-spatial frequency components compared to the far-field light, it is a key parameter to overcome resolution limit in optical microscopy. In this thesis, we developed two different types of super-resolution microscopy. One is a near-field scanning optical microscopy (NSOM). NSOM can provide sub-diffraction resolution by directly detecting the near-field light using a metal coated tip with a small aperture. To obtain a super-resolution image from light-emitting diode (LED), a home-built NSOM system was combined with spectroscopy. By applying correlation analysis, LED samples were quantitatively analyzed in sub-diffraction scale. The other one is super-resolution optical fluctuation imaging (SOFI). SOFI enables to provide super-resolution images by estimating high-spatial frequency components in near-field light using statistical analysis such as correlation from temporal optical fluctuations of fluorophores referred to as blinking. However, because SOFI utilizes intrinsic blinking property of fluorophores, it is only applicable to specific types of fluorophore. To overcome this limitation, we developed a new method by combining SOFI with speckle pattern illumination (S-SOFI). Because S-SOFI uses controllable optical fluctuation induced by illumination, it is applicable for any types of fluorophores. Applying S-SOFI, we theoretically and experimentally demonstrated its ability of super-resolution imaging. Furthermore, S-SOFI was combined with near-field speckle pattern (NS-SOFI) to further improve the resolution of S-SOFI. Using NS-SOFI, we experimentally obtained an image with improved resolution compared to S-SOFI.