Abstract: Abstract: In order to study how shear strength and its change in unsaturated soils modified by hydrophobic agent, sandy soils hydrophobized by dimethyldichlorosilane (DMDCS) were prepared. The hydrophobic soils with different DMDCS volume by soil mass, water content and clay content were obtained respectively. Mixtures of sandy soils (DMDCS volume by soil mass were 2 mL:100 g and 3 mL:100 g, i.e., 2% and 3%, the same below) with 3%, 5%, 7%, 11% and 13% water content were prepared respectively. Meanwhile, the mixtures of clay and sandy soils (3% DMDCS) according to the clay mass fraction of 0, 5%, 10%, 15%, 20%, 25%, 30%, 40% and 50% were prepared respectively. Then the unsaturated direct shear tests were carried out by unsaturated direct shear apparatus. Four samples in each group were prepared under the normal pressure in 100, 200, 300 and 400 kPa, respectively. The horizontal shearing ratio was 0.8 mm/min and the shearing test lasted 6 min. The results showed that: 1) The mixtures of sandy soils and DMDCS presented extreme hydrophobicity. 2) The water repelling of mixtures of clay and sandy soils was affected by both DMDCS and clay content. The water repelling of mixtures was degraded with the increasing of clay content, and improved with the increasing of DMDCS volume by soil mass. 3) The Mohr-Coulomb strength criterion could be adopted to describe the shear strength of hydrophobic sandy soils with different DMDCS and water content. With the increasing of DMDCS volume by soil mass, the shear strength index of the hydrophobic soil was decreased to different extents. The cohesive force showed a steep-drop shape from soil without DMDCS addition to that with 1% DMDCS and a slow decline up to stabilize finally. The effect of DMDCS on the cohesive force of modified sandy soils was much greater than on the internal friction angle. The hydrophobic soils' cohesion with the increasing of water content was decreased gradually, while the internal friction angle was increased firstly and decreased finally. The water content had a much effect on the shear strength of the hydrophobic sandy soils. The shear strength index were decreased with the increasing of DMDCS volume by soil mass: the cohesion was declined rapidly from 19.6 kPa (no DMDCS) to 10.4 kPa (1% DMDCS) and slowly from 10.4 kPa (1% DMDCS) to 9.3 kPa (3% DMDCS). The internal friction angle was declined slowly from 16.2° (no DMDCS) to 11.8° (3% DMDCS). The cohesion was decreased with the increasing of water content: the cohesion was declined from 15.5 kPa (3% water content) to 7.5 kPa (13% water content) with 2% DMDCS and from 15 kPa (3% water content) to 5.2 kPa (13% water content) with 3% DMDCS respectively; the internal friction angle was increased from 12.7° (3% water content) to 15.2° (11% water content) with 2% DMDCS and from 11.7° (3% water content) to 14.6° (7% water content) with 3% DMDCS, then decreased to 13.9° (13% water content) with 2% DMDCS and 11.9° (13% water content) with 3% DMDCS respectively. 4) The Mohr-Coulomb strength criterion could also be adopted to describe the shear strength of hydrophobic mixed soils with different clay content. The cohesion was increased remarkably with the increasing of clay content. The cohesion of sandy soils without clay was 9.3 kPa, then suddenly increased to 27.2 kPa and 55.1 kPa mixed with 5% and 50% clay mass fraction, respectively. The internal friction angle was increased from 12.1° (5% clay mass fraction) to 16.2° (15% clay mass fraction), and then decreased to 9.7° (50% clay mass fraction). All the analysis could be beneficial to analyze the shear strength of hydrophobized soils and apply in engineering.