An experimental design with factorial planning was used for the immobilization of the enzyme cyclodextringlycosyltransferase (CGTase) from Bacillus firmus (strain no. 37) to select the best combination of support, method of immobilization, and conditions that gives primarily higher average values for the specific immobilized enzyme activity, and secondarily, higher average values for the percentage of protein fixation. The experimental design factors were as follows: supports-controlled-pore silica, chitosan, and alumina; immobilization methods-adsorption, and two covalent bonding methods, either with gamma-aminopropyltriethoxysilane or hexamethylenediamine (HEMDA); conditions-7degreesC without agitation and 26degreesC with stirring. The best combination of factors that lead to higher average values of the response variables was obtained with immobilization of CGTase in silica with HEMDA at 7degreesC. However, immobilization in chitosan at 7degreesC gave the highest immobilized CGTase specific activity, 0.25 mumole of beta-CD/(min.mg protein). Physical adsorption gave low specific enzyme activities, and, in general, a high load of enzyme leads to lower specific enzyme activity.