In-vivo information of myelin content is desirable for studying many brain diseases and injuries which damage myelin. Myelin water imaging (MWI) is a validated and quantitative MR method to myelin. However, the data post-processing of MWI is mathematically complex and computationally demanding. The analysis typically takes several hours for a whole brain analysis, which limits its clinical applications. Our objective was to train a neural network as an alternative method for the MWI data analysis. We found this novel approach can accelerate MWI data analysis by over 150 times.