Objective measurement of meat tenderness has been a topic of concern for palatability evaluation. In this study, a real-time ultrasonic B-mode imaging method was used for measuring beef palatability attributes such as juiciness, muscle fiber tenderness, connective tissue amount, overall tenderness, flavor intensity, and percent total collagen noninvasively. A temporal averaging image enhancement method was used for image analysis. Ultrasonic image intensity, fractal dimension, attenuation, and statistical gray-tone spatial-dependence matrix image texture measurement were analyzed. The contrast of the textural feature was the most correlated parameter with palatability attributes. The longitudinal scanning method was better for juiciness, muscle fiber tenderness, flavor intensity, and percent soluble collagen, whereas, the cross-sectional method was better for connective tissue, overall tenderness. The multivariate linear regression models were developed as a function of textural features and image intensity parameters. The determinant coefficients of regression models were for juiciness (R2 equals .97), for percent total collagen (R2 equals .88), for flavor intensity (R2 equals .75), for muscle fiber tenderness (R2 equals .55), and for overall tenderness (R2 equals .49), respectively.