In this study, the water state and distribution of Chinese dried noodles during drying process was investigated by low-field nuclear magnetic resonance (LF-NMR). Transverse relaxation times (T-2) were achieved by LF-NMR coupled with a 0.5 T permanent magnet equivalent to a proton resonance frequency of 21 MHz at 32 degrees C. Three populations of water state can be distinguished: strongly bound water (T-21, 0.04 -0.40 ms; A(21), 0.25-19.08%), weakly bound water (T-22, 0.96-534 ms; A(22), 80.81-98.44%), and free water (T-23, 74.50-266.47 ms; A(23), 0.11-1.61%). During the drying process, the transverse relaxation time of all water states followed decreasing trend. Initially, the moisture content decreased faster from the edges; however, moisture migrated rapidly from the central part during 90-180min. The moisture gradient disappeared after drying for 240min. Besides, a special method as "Oil immersion method" was introduced to address poor signal to noise ratio in samples when moisture content was low during the drying process. High signal to noise ratio was successfully achieved by this method. (C) 2018 Published by Elsevier Ltd.