The glycaemic index (GI) of commonly consumed bakery products supplemented with increasing levels of coconut (Cocos nucifera) flour was determined in ten normal and ten diabetic subjects. Using a randomized crossover design, the control and test foods were fed in random order on separate occasions after an overnight fast. Blood samples were collected through finger prick before and after feeding and were analysed for glucose levels using a clinical chemistry analyser. The significantly low-GI (< 60) foods investigated were: macaroons (GI 45·7 (sem 3·0)) and carrot cake (GI 51·8 (sem 3·3)), with 200–250 g coconut flour/kg (P<0·05). The test foods with 150 g coconut flour/kg had GI ranging from 61·3 to 71·4. Among the test foods, pan de sal (GI 87·2 (sem 5·5)) and multigrain loaf (GI 85·2 (sem 6·8)) gave significantly higher GI with 50 and 100 g coconut flour/kg respectively (P<0·05). On the other hand, granola bar and cinnamon bread with 50 and 100 g coconut flour/kg respectively gave a GI ranging from 62·7 to 71·6 and did not differ significantly from the test foods with 150 g coconut flour/kg (P < 0·05). A very strong negative correlation (r– 0·85, n 11, P < 0·005) was observed between the GI and dietary fibre content of the test foods supplemented with coconut flour. In conclusion, the GI of coconut flour-supplemented foods decreased with increasing levels of coconut flour and this may be due to its high dietary fibre content. The results of the present study may form a scientific basis for the development of coconut flour as a functional food. However, the fat content of coconut flour-supplemented food should always be considered to optimize the functionality of coconut fibre in the proper control and management of diabetes mellitus.