The beef quality grade is greatly affected by visible fat content. Especially, in Japanese black (Wagyu) cattle, high fat content is typically valued highly. In this paper, we describe the feasibility of beef evaluation by visualizing fat characteristics using near-infrared (NIR) multispectral imaging. An intact raw beef cut from Wagyu cattle was used as an evaluation target. The content of fat and fatty acid, such as the total saturated fatty acid (SFA) content, the total unsaturated fatty acid (UFA) content, myristic acid (C14:0), palmitic acid (C16:0), stearic acid (C18:0), myristoleic acid (C14:1), palmitoleic acid (C16:1), oleic acid (C18:1), and linoleic acid (C18:2) were estimated and visualized. The total SFA content was calculated as the sum of myristic acid, palmitic acid, and stearic acid. Also, the total UFA content was calculated as the sum of myristoleic acid, palmitoleic acid, oleic acid, and linoleic acid. Reference values for the fat content and fatty acid composition were determined by conventional physical and chemical methods. The fatty acid composition was determined from the extracted lipids by Folch's method, by gas chromatography (GC) using its methyl ester. The fat content was determined by using the Gerhardt SOXTHERM. The NIR multispectral images of the sample were acquired by using the SPECIM Spectral Camera SWIR. It works in the wavelength range of 970-2500 nm with 6.3 nm of bandwidth at 320 pixels resolution in spatial domain. The absorbance spectra of each pixel calculated from pixel intensity of subject and reference white standard was used for constructing the prediction model. In total, 33 samples from various parts of the 2 head of Wagyu cattle were measured. Calibrations were performed by a partial least squares (PLS) regression using mean extracted spectra from each individual sample, limited wavelength range from 1000 to 2300 nm. The coefficients of determination (R2) were between 0.68 and 0.87. The ranks by evaluation index (EI) were “B (high accuracy)” and “C (slightly high)”. The ratios of the standard error of prediction to the standard deviation (RPD) were between 1.74 and 2.74. These results indicate a sufficient feasibility of the prediction except for myristoleic acid content. The visualizations, which show the spatial distribution of fatty acid content, were performed by applying the model to predict the content of each pixel.