Pathological changes of the body have been observed to change the mechanical properties of soft tissue types which can be imaged by ultrasound elastography. Though initial clinical results using ultrasound elastography in detection of tumors are promising, quantification of signal to noise ratio, resolution and strain image patterns are the best achieved under a controlled study using phantoms with similar biomechanical properties of normal and abnormal tissues. The purpose of this work is to characterize the biomechanical properties of agar based tissue mimicking phantoms by varying the agar concentration from 1.7 to 6.6% by weight and identify the optimum property to be used in classification of cancerous tissues. We performed quasi-static uniaxial compression test under a strain rate of 0.5. mm/min up to 15% strain and measured Young's modulus of phantom samples which are from 50. kPa to 450. kPa. Phantoms show nonlinear stress-strain characteristics at finite strain which were characterized using hyperelastic parameters by fitting Neo-Hookean, Mooney Rivlin, Ogden and Veronda Westmann models. We also investigated viscoelastic parameters of the samples by conducting oscillatory shear rheometry at various precompression levels (2-5%). Loss modulus values are always less than storage modulus which represents the behavior of soft tissues. The increase in agar concentration increases the shear modulus of the samples as well as decreases the linear viscoelastic region. The results suggest that dynamic shear modul are more promising than linear and nonlinear elastic modul in differentiation of various classes of abnormal tissues. © 2014 Elsevier Ltd.