Background The aim of this study is to identify cutoff values for muscle ultrasound (US) to be used in Global Leadership Initiative on Malnutrition (GLIM) criteria, and to define the effect of reduced muscle mass assessment on malnutrition prevalence at hospital admission. Methods A total of 118 inpatients were enrolled in this cross-sectional study. Six different muscles were evaluated by US. Following defining thresholds for muscle US to predict low muscle mass measured by bioelectrical impedance analysis, malnutrition was diagnosed by GLIM criteria with seven approaches, including calf circumference, mid-upper arm circumference (MAC), handgrip strength (HGS), skeletal muscle index (SMI), rectus femoris (RF) muscle thickness, and cross-sectional area (CSA) in addition to without using the reduced muscle mass criterion. Results The median age of patients was 64 (18-93) years, 55.9% were female. RF muscle thickness had moderate positive correlations with both HGS (r = 0.572) and SMI (r = 0.405). RF CSA had moderate correlation with HGS (r = 0.567) and low correlation with SMI (r = 0.389). The cutoff thresholds were 11.3 mm (area under the curve [AUC] = 0.835) and 17 mm (AUC = 0.737) for RF muscle thickness and 4 cm(2) (AUC = 0.937) and 7.2 cm(2) (AUC = 0.755) for RF CSA in females and males, respectively. Without using the reduced muscle mass criterion, malnutrition prevalence was 46.6%; otherwise, it ranged from 47.5% (using MAC) to 65.2% (using HGS). Conclusions Muscle US may be used in GLIM criteria. However, muscle US needs a standard measurement technique and specific cutoff values in future studies.