The similarity model test is one of the important means to study the engineering properties of red-bed soft rock; the test result reliability depends on the correct selection and rational proportion of a series of similar materials. In view of the softening characteristics of the saturated soft rock

a series of mixture ratio tests on similar material with orthogonal design method is conducted by adding the quicklime into the traditional similar materials. Four influential factors are considered in all the tests

including the ratio of iron pore powder to barite powder to aggregates

the content and ratio of iron pore powder to iron pore powder plus barite powder

the content of gypsum

and the content of quicklime. Each factor is classified into 5 levels. In total

610 specimens in 25 groups were prepared. For each group of specimens

the physico-mechanical indices

such as softening coefficient

density

unconfined uniaxial compressive strength

tensile strength

cohesive force

etc. were tested. On the basis of these test specimens

the influential factors of the concerned parameters were analyzed; meanwhile the applicability of the ratio result were examined. At last

the similar materials of three typical soft rock specimens were prepared to test the softening coefficient; and a comparison analysis was also carried out with the real soft rock test. The test results reveal that: 1) the physico-mechanical properties of the similar material fabricated according to the similarity ratio proposed in this paper may vary considerably

and this can meet the requirement of the geomechanical model test of soft rock; 2) the physico-mechanical properties of the similar material demonstrate an evident variation law that facilitates the identification of the optimum mixture ratio for any soft rock material; 3) the softening coefficients of the saturated similar material

in the three comparative tests

are consistent with the real soft rock to a great extent.