This dissertation focuses on the development of analytical methods based on Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) and their applications for separation and structural characterization of oligosaccharides. Porous graphitized carbon liquid chromatography (PGC-LC), gated-trapped ion mobility spectrometry (Gated-TIMS), and electronic excitation dissociation tandem mass spectrometry (EED MS/MS) are three essential techniques employed here. First, the EED method was optimized to generate more informative glycan tandem mass spectra for accurate structural analysis. Glycans were reduced and permethylated or labeled with a reducing-end fixed charge to increase sensitivity, avoid gas-phase structural rearrangement, and facilitate spectral interpretation. EED of glycans produced nearly complete series of Z-, Y- and 1,5X-ions, that appear in the spectra as triplets with characteristic spacing, thus facilitating accurate determination of the glycan topology. Additional radical-driven dissociation pathways were identified, from which different types of linkage-diagnostic ions (cross-ring, secondary, or internal fragments) were generated. The results demonstrated that linkage analysis can be accomplished by utilizing one or a combination of several linkage-diagnostic fragments. EED MS/MS was then implemented, in conjunction with PGC-LC or Gated-TIMS, for on-line separation and characterization of complex mixtures of glycans. These two methods were successfully applied for high-throughput and detailed structural analysis of N-glycans released from human serum, O-glycans released from bovine submaxillary mucin and free oligosaccharides. The performance of these methods was tested and improved through analysis of different types of glycans from a variety of biological sources. Finally, in collaboration with bioinformaticians, a spectral interpretation algorithm, GlycoDeNovo, has been developed for automated and de novo glycan topology reconstruction from their tandem mass spectra. A large number of EED tandem spectra of glycan standards generated in house were used as the training dataset to establish appropriate IonClassifiers for candidate ranking. GlycoDeNovo is capable of identifying correct topologies from MS/MS spectra of glycans in different derivatized forms. Several aspects of this collaborative project were covered in this thesis, including glycan derivatization, data acquisition and manual spectral interpretation to guide the development and evaluate the performance of the automated approach. In this thesis research, integrated approaches utilizing PGC-LC–EED-MS/MS and Gated-TIMS–EED-MS/MS, and the appropriate bioinformatics software, have been established for structural analysis of glycan mixtures. They hold great potential for comprehensive, automated, and de novo glycome characterization.