Optimization of culture conditions for xylitol production from xylose by Candida parapsilosis KFCC 10875 was performed. Optimum medium compositions were determined to be 50 g/l xylose, 5 g/l yeast extract, 5 g/l $(NH_4)_2SO_4$, 5 g/l $KH_2PO_4$, and 0.2 g/l $MgSO_4$·$7H2_O$ and optimum culture conditions were pH 4.5-5.5, 30℃, 200 rpm in a 5 liter jar fermentor. A final xylitol concentration of 36 g/l was obtained from 50 g/l of xylose after 54 h under the optimum culture conditions. The effect of byproducts involving glucose, acetic acid and arabinose, which were derived from corn cob hemicellulose hydrolysate, on xylitol fermentation was investigated. With increasing the ratio of glucose to xylose, xylitol production decreased, ethanol and glycerol production increased. Glucose was contributed to only cell growth but not to xylitol production. Cell growth, xylose consumption, and xylitol production decreased when acetic acid concentration was increased. However, the xylitol yield from xylose and specific production rate of xylitol were maximum at 1.0 g/l acetic acid. When the arabinose was added in xylose medium, xylitol production decreased but ethanol production increased. The total products (xylitol and ethanol) yield was constant regardless of the arabinose concentration, suggesting that arabinose shifted the carbon flow from xylitol to ethanol. When corn cob hemicellulose hydrolysate was used as a substrate for xylitol production, cell growth and ethanol production increased due to the stimulating effects of glucose and arabinose, whereas xylitol production increased due to the stimulating effect of a low concentration of acetic acid. Improvement of xylitol productivity was performed by controlling oxygen supply. The maximum production of xylitol from 50 g/l xylose was achieved at a kLa of 45 $h^{-1}$, a dissolved oxygen (DO) concentration of 1.5%, a redox potential of 100 mV, indicating that the oxygen supply must be carefully controlled for efficient xylitol production from xylose. The redox potential was found to be useful for monitoring the DO level in culture medium, especially when the DO level was low. A two-step xylitol production strategy was devised, with first-step involving rapid production of cells under well-aerated conditions, and second-step involving cultivation with reduced aeration such that the culture redox potential was 100 mV. By using two-step fed-batch culture, a final xylitol concentration of 210 g/l which corresponded to a xylitol yield of 70%, was obtained in a 5 l fermentor from 300 g/l xylose in 66 h and a final xylitol concentration of 180 g/l was obtained from a culture medium totally containing 254.5 g/l xylose in a 3,000 l pilot scale fermentor after 77 h fermentation.