An NMR spectrometer over 1 GHz requires the contribution of high-c superconductors (HTS). However, a persistent-mode operation with HTS cannot satisfy the field stability of 0.01 ppm/h at present. This is a great barrier for applying HTS to an NMR magnet. To overcome this problem, a new project was undertaken in Japan in October 2006. In the course of the project, we will develop a highly stabilized power supply, field-compensation methods, and measurement techniques that allow a certain field fluctuation. By integrating them, the feasibility of HTS to NMR will be demonstrated. We performed a long-term operation of a 600 MHz NMR magnet in the driven-mode. Allowable field fluctuation of the existing internal lock system for solution NMR was evaluated by a model experiment. As the next step, the innermost Nb 3 Sn coil of the 600 MHz NMR magnet will be replaced with a Bi-2223 coil, and the field homogeneity, as well as the field stability, will be evaluated. In the final step of the project, the replacement of the innermost coil of the existing 920 MHz NMR magnet will be planned. The targeting field is 24.7 T (1.05 GHz for 1 H NMR resonance frequency). The solid-state NMR on 17 O nuclei in a labeled peptide will be demonstrated using a magic angle spinning probe; the probe has a 1 H decoupling frequency of 1.05 GHz.Index Terms-Field stability, high magnetic field, high temperature superconductor, NMR spectrometer.