We compared the in vitro/in vivo properties of theophylline between two sustained-release preparations, which are administered once a day. Tablet A is a swelling/disintegration-type matrix tablet consisting of hydrophobic wax granules and hydrophilic polymer granules (cluster tablets). Tablet B is a matrix tablet consisting of hydrophilic polymer granules. We conducted a dissolution test with JPXIV in vitro, and compared the results between the two preparations. Neither pH nor agitation intensity influenced these preparations. After they were immersed in oleic acid, there were no marked changes in the dissolution properties in the dissolution test. After administration of Tablets A and B containing theophylline at 200mg to fasted dogs, we compared plasma level profiles of theophylline. The mean plasma level of theophylline gradually increased to a maximum (7.17microg/mL) 4h after administration of Tablet A. After administration of Tablet B, a similar finding was noted, with a maximum of 6.09microg/mL. Tablet B showed a higher coefficient of variation (CV) for the plasma level at each point. Subsequently, we administered two tablets of preparations A and B containing theophylline at 200mg to healthy volunteers who had not been fasted, and compared plasma level concentration of theophylline. The mean plasma level of theophylline gradually increased to a maximum (6.09microg/mL) 12h after administration of Tablet A, but then decreased, with a half-life of 9.10h. After administration of Tablet B, a similar finding was noted, with a maximum of 7.87microg/mL and a half-life of 7.76h. Tablet A showed a significantly higher plasma concentration 1 and 2h after administration; however, there were no significant differences at other points. The C(max) of Tablet B was significantly higher than that of Tablet A. However, there were no significant differences in other pharmacokinetic parameters between the two preparations. The T(max) of Tablet A was 10-12h after administration, relatively constant. However, that of Tablet B was 10-18h after administration. The CV for T(max) was 9.8% for Tablet A and 22.0% for Tablet B. After administration of Tablet B, the plasma level of theophylline varied at each point. Based on these results, inter-subject variations after administration of Tablet A may be less marked than those after administration of Tablet B. It is concluded that, the cluster tablets A developed in this study showed significantly less inter-subject variation of theophylline plasma levels than the conventional matrix tablets B.