The human body contains a variety of tissue species with short T2 ranging from a few microseconds to hundreds of microseconds. Detection and quantification of these short-T2 species is of considerable clinical and scientific interest. Cortical bone water and myelin are two of the most important tissue constituents. Quantification of cortical bone water concentration allows for indirect estimation of bone pore volume and noninvasive assessment of bone quality. Myelin is essential for the proper functioning of the central nervous system (CNS). Direct assessment of myelin would reveal CNS abnormalities and enhance our understanding of neurological diseases. However, conventional MRI with echo times of several milliseconds or longer is unable to detect these short-lived MR signals. Recent advances in MRI technology and hardware have enabled development of a number of short-T2 imaging techniques, key among which are ultra-short echo time (UTE) imaging, zero echo time (ZTE) imaging, and sweep imaging with Fourier transform (SWIFT). While these pulse sequences are able to detect short-T2 species, they still suffer from signal interference between different T2 tissue constituents, image artifacts and excessive scan time. These are primary technical hurdles for application to whole-body clinical scanners. In this thesis research, new MRI techniques for improving short-T2 tissue imaging have been developed to address these challenges with a focus on direct detection and quantification of cortical bone water and myelin on a clinical MRI scanner. The first focus of this research was to optimize long-T2 suppression in UTE imaging. Saturation and adiabatic RF pulses were designed to achieve maximum long-T2 suppression while maximizing the signal from short-T2 species. The imaging protocols were optimized by Bloch equation simulations and were validated using phantom and in vivo experiments. The results show excellent short-T2 contrast with these optimized pulse sequences. The problem of blurring artifacts resulting from the inhomogeneous excitation profile of the rectangular pulses in ZTE imaging was addressed. The proposed approach involves quadratic phase-modulated RF excitation and iterative solution of an inverse problem formulated from the signal model of ZTE imaging and is shown to effectively remove the image artifacts. Subsequently image acquisition efficiency was improved in order to attain clinically-feasible scan times. To accelerate the acquisition speed in UTE and ZTE imaging, compressed sensing was applied with a hybrid 3D UTE sequence. Further, the pulse sequence and reconstruction procedure were modified to enable anisotropic field-of-view shape conforming to the geometry of the elongated imaged object. These enhanced acquisition techniques were applied to the detection and quantification of cortical bone water. A new biomarker, the suppression ratio (a ratio image derived from two UTE images, one without and the other with long-T2 suppression), was conceived as a surrogate measure of cortical bone porosity. Experimental data suggest the suppression ratio may be a more direct measure of porosity than previously measured total bone water concentration. Lastly, the feasibility of directly detecting and quantifying spatially-resolved myelin concentration with a clinical imager was explored, both theoretically and experimentally. Bloch equation simulations were conducted to investigate the intrinsic image resolution and the fraction of detectable myelin signal under current scanner hardware constraints. The feasibility of quantitative ZTE imaging of myelin extract and lamb spinal cord at 3T was demonstrated. The technological advances achieved in this dissertation research may facilitate translation of short-T2 MRI methods from the laboratory to the clinic.