The water quality improvement of golf course ponds, as representative stagnant stream channels, was evaluated by applying a biological activated carbon (BAC) process composed of four consecutive activated carbon reactors. The study was performed from autumn to winter in order to evaluate the feasibility of the BAC process under low temperature conditions. In the study, water quality of pond A (target pond) and pond B (reference pond) were monitored. Pond water was pumped into the BAC process, and was then returned to the pond after treatment. The optimal conditions were determined to be 2 hr of empty bed contact time (EBCT) at a temperature above $4^{circ}C$, in which improvements of chemical oxygen demand (COD), total nitrogen (TN) and total phosphorus (TP) of pond A compared to pond B were 3.62%, 3.48% and 1.81%, respectively. On the other hand, as the temperature was below $4^{circ}C$, some degree of water quality improvement was achieved even when EBCT were 1 or 0.5 hr, suggesting that the BAC process can be successfully applied for the improvement of pond water quality in winter months. The values of biomass concentration and microorganism activity in each condition were highest where 2 hr of EBCT was applied at a temperature above $4^{circ}C$, but values were similar throughout all treatment conditions, and thus, adsorption is considered to be the dominant factor affecting process efficiency. From the denaturing gel gradient electrophoresis (DGGE) results, no significant differences were observed among the activated carbon reactors, suggesting that the number of reactors in the system could be decreased for a more compact application of the system.