Scanning Near-field Optical Microscopy (SNOM) technique enables to overcome Abbe diffraction limit of far-field optics as well as to obtain simultaneously optical and topographical images. While the optical resolution of the method is limited by the aperture size and is typically 50 - 100 nm, an excellent spatial resolution in a topography channel can be realized. Naturally, we need a convenient and precise method to control the distance between the tip and sample for the successful operation of any SNOM device. Nowadays, the most popular method of the SNOM tip-sample distance control is the shear force - based feedback employing a glass fiber attached to the quartz tuning fork (TF). However, the shear-force distance control method is far from the ideal one. The crosstalk between optical and topographical image can warp the results. The forces between the tip and sample are high and in many cases might be destructive. We report the realization of a new approach to the problem: bent sharpened glass optical fibers with carefully controlled sizes of the bent part and the radius of the curvature of the bending were prepared and experimentally exploited as SNOM probes. Detailed analysis of fiber vibration modes shows that realization of truly tapping mode of the probe dithering requires extreme caution. In case of using the second resonance, mode probes vibrate mostly in the shear-force mode unless the bending radius is rather small (0.3 mm) and the probe's tip is short. The probes having these characteristics were prepared and attached to the TF in the double resonance conditions, which enables to achieve a significant quality factor of the sensor. Another common problem of most aperture SNOMs is the fragility of the tip. We proposed and realized the use of different plastic fibers to solve this problem. These fibers look very promising for the use as SNOM probes and are characterized by much less fragility (compare glass and plastic) and greater ease of the tip preparation. For such preparation, hazardous treating with hydrofluoric acid, which remains the most popular approach to prepare SNOM probes from the glass fibers, can be entirely avoided. Fluorescence Resonance Energy Transfer (FRET) is one of the most promising ways to improve the spatial resolution of the SNOM, and the central part of the Thesis is devoted to the elaboration of FRET SNOM. The idea is to use a donor (acceptor) nanoparticle/molecule as local fluorescence center attached to the tip and measure the fluorescence induced by it in the sample (or vice versa) due to the FRET. Ten years ago, this idea was realized at the single molecule level with CdSe nanocrystals and appropriate dye molecules. Despite the high spatial resolution (better than 20 nm) attained in this experiment, it remains an isolated one, and this is for a valid reason: albeit rather large, the photostability of dye molecules and semiconductor nanocrystals still enables to use a single fluorescence center exploiting for imaging only a few minutes at best. Fluorescent centers with high photostability should be used to overcome this problem. Earlier, claimed to be very photostable and bright NV color centers in nanodiamond crystals were proposed. In the Thesis, we show that such a system is not suitable to realize a single fluorescent center FRET SNOM method. We propose to use specific rare-earth ions in crystals to achieve the goal, in particular, LuBO3(Tb) micro- and nanocrystals.