Optical measurement is one of the most important means of Space Object Identification. Due to the long observation distance, space objects in the field-of-view usually turn out to be unresolved points. Compared with orbit determination and photometry measurements, spectral measurements provide useful information in the wavelength dimension, which greatly improves the ability of material discrimination. This ability has great potential since it can provide reliable evidences for space object characterization. This paper introduces some typical results of space objects spectral observation and characterization in recent years, including multi-band photometry measurement, hyperspectral measurement, laboratorial spectral measurement, and characteristics modeling and simulation. Multi-band photometry measurement can bring low resolution spectral data, which is a common method for space object classification in large area. Hyperspectral measurement focuses on detailed spectral structure in the reflected light, from whom material composition of specific objects can be derived. Laboratorial spectral measurement can simulate the process of real observation under controlled conditions, while providing database of materials’spectral data. Characteristics modeling and simulation is used to generalize the attribute data to study the spectrum changing process. By analyzing domestic and overseas experiences, the capacity and weakness of current researches as well as several ideas of future works are discussed, offering a reference for future studies.