BACKGROUND Muscle mass is often mentioned not to reflect muscle strength. For muscle mass assessment skeletal muscle index (SMI) is often used. We have reported that dual-energy X-ray absorptiometry (DXA)-derived SMI does not change with age in women, whereas the cross-sectional muscle area (CSMA) derived from computed tomography (CT) does. OBJECTIVE The present study aimed to compare CT and DXA for the assessment of muscle tissue. METHODS Cross-sectional study in the local residents. METHODS A total of 1818 subjects (age 40-89 years) randomly selected from community dwellers underwent CT examination of the right mid-thigh to measure the cross-sectional muscle area (CSMA). Skeletal muscle mass (SMM) was measured by DXA. The subjects performed physical function tests such as grip strength, knee extension strength, leg extension strength, and gait speed. The correlation between CT-derived CSMA and DXA-derived SMM along with their association with physical function was examined. RESULTS After controlling for related factors, the partial correlation coefficient of muscle cross-sectional area (CSA) with physical function was larger than that of DXA-derived SMM for gait speed in men (p=0.002) and knee extension strength in women (p=0.03). The partial correlation coefficient of quadriceps (Qc) CSA with physical function was larger than that of DXA-derived SMM for leg extension power in both sexes (p=0.01), gait speed in men (p<0.001), and knee extension strength in women (p<0.001). CONCLUSIONS Mid-thigh CT-derived CSMA, especially Qc CSA, showed significant associations with grip strength, knee extension strength, and leg extension power, which were equal to or stronger than those of DXA-derived SMM in community-dwelling middle-aged and older Japanese people. The mid-thigh CSMA may be a predictor of mobility disability, and is considered to be useful in the diagnosis of sarcopenia.