In the autumn of 2000 and spring of 2001, field surveys were conducted to estimate the effectiveness of artificial reefs (type cube, $2{ imes}2{ imes}2m^3$) that were established in the four islands of Bangnyeong, Socheong, Daeyeonpyeong and Ganghwa in Kyonggi Bay, the west coast of Korea during 1995 and 1996. The condition of reefs was examined through SCUBA diving and a side-scan sonar. Much of the reefs in Daeyeonpyeong and Ganghwa area were buried in bottom sediment. Despite an intensive search in Bangnyeong area, even a cluster of reefs was not found and most of them seemed to be buried by sand waves. Thus an appropriate investigation on the sediment transport should be included in pre-assessment for the expected performance and protection of artificial reefs. Distribution of average CPUE in natural fishing ground (control) was estimated by bootstrapping simulation and possible comparison of CPUE between control and reef areas (treatment) were made in Bangnyeong and Socheong (Experiment I). Positive reef effect was detected in Socheong but CPUE of treatment in Bangnyeong was varied between or lower than the 99% CPUE confidence intervals of the control. Control/treatment abundance and biomass of fishes and invertebrates were tested by paired t-test and sign test (Experiment II). Only four cases among 22 showed significant positive effect. Based on the results, the cube artificial reef in Socheong was inferred as an affirmative one. Floor type was hypothesized to be one of the probable agents in determining the effectiveness of artificial reefs.