In Japan, satoyama has been often abandoned due to low and ageing population, resulting in deterioration of ecosystem service and biodiversity. In contrast, restoration practice seems to improve them, which is to be studied quantitatively. We aim to monitor the change in pollination system at the onset of restoration of satoyama abandoned 40 years ago. Sixteen pairs of restored and control plots (both 2 x 2 m² ) were established in the study site in Kanazawa. We constructed a paddy inside each restored plot and weeded the bank, while control plots were left untouched. For these plots, number of species and percentage coverage of flowering plants were recorded and flower-visiting bees were collected using sweeping net 3 times per month from May to October, 2013. Results indicate that (1) bee diversity and flowering plant-bee linkage parameters were higher in control than restored plots; (2) flowering plant diversity parameters were variable among plots; (3) both plots were different in seasonality of flowering plants and bees; (4) plant-bee linkages in control plots were more diverse than in restored plots because restoration was still at the beginning. Further restoration process must be followed in this field.