Abstract: This paper reports on the orthogonal experiments of treating conditions with polyethylene glycol (PEG) for the wood of Chinese pine (Pinus massoniana Lamb.), ash (Fraxinus mandshurica Rupr.) and basswood (Tilia amurensis Rupr.).The molecular weight of polyethylene glycol used for the experiments was 1000. Orthogonal design of the experimentsThe ratio between the dry weights of the untreated wood and the polymer of PEG solution was 1.5. Soaking and drying were carried out at room temperature.The data to be investigated were: 1) the total take-up of dry polymer per unit dry weight of the untreated wood, 2) the weight increase per unit dry weight of the untreated wood of the moisture-contained sp cimens in the polymrr solution, 3) the diffusion coefficient during soaking periods, 4) the maximum radial and tangential shrinkage percentage from ths soaking to the ovendried condition and 5) the radial and taengntial anti-shrinkage efficiencies (ASE). Variance analyres were made of all this.For air-dried cross sections with a thickness of about 0.75 cm along the grain, the PEG solution could impregnate the cell of wood through the capillary attraction of the cell cavity and cell wall and the mean tangential ASE of pine and ash reached 78.26-90.20 percent and 48.11-64.83 percent respectively. However, for wood in lumber length, capillary action is very weak. Thus in commercial applieation, the treatment of wood with PEG chiegly relies on diffusion; the initial moisture content (MC) of wood becomes the most important factor affecting the dimensional stabilization of treated wood. An initial MC of 80 and 120 percent arc considered to be feasible for pine and ash respectively.During soaking, all the samples exhibited a little shrinkage cxc.pt the air-dried. This shows that the diffusion not only means the equal volume exchange between polymer and water, but also a little dehydration of the cell wall.