Arrangements of elementary soil particles during soil formation in combination with biological activity by plant roots and earthworms may yield anisotropy of the non-tilled subsoil pore system. Soil compaction by agricultural machinery is known to affect the soil pore characteristics but only few studies have addressed its effect on the soil pore anisotropy. We addressed anisotropy in two long-term field compaction experiments in Finland (clay) and Sweden (sandy loam) that were established in 1980 and 1995, respectively. In 2009/2010, soil cores of 100 cm³ were collected vertically and horizontally at 0.3, 0.5, 0.7 and 0.9 m depths (the two lower depths only in Sweden) in two treatments (compacted and control). Water retention, air permeability (k_a) and gas diffusivity (D_s/D_o) were determined in the laboratory. We defined an anisotropy factor (AF) as the ratio of a certain soil property measured in the horizontal direction to the same property measured in vertical direction. For both soils and both treatments, k_a at -100 hPa was higher in vertical direction than in horizontal direction (AF < 1). Anisotropy was also observed for D_s/D_o in all depths, with AF in the range 0.77 – 1.60 and 0.12 – 0.34 in the sandy loam soil and clay soil, respectively. AF was closer to 1 for the compacted treatments for both soils. We applied Ball’s tube model for the pore system and calculated the number of soil pores in a given soil cross-sectional area and the tube-equivalent pore diameter. Vertical cores had fewer pores with larger diameter than horizontal cores. The slope of log k_a = f(log (air-filled porosity)) indicated that for the sandy loam soil, the rate of opening of continuous air-filled pathways was greater in the vertical than in the horizontal direction. Soil compaction generally reduced the anisotropy of the soil pore system.