The main objective of this study was to observe the sludge accumulation pattern of an experimental, covered, anaerobic pond treating municipal wastewater under Mediterranean climatic conditions throughout a 2-year operational period (1999-2000) in order to form a seasonal sludge accumulation model which may be used to predict the required desludging time, not only of the particular anaerobic pond used in the study, but also for other types of anaerobic ponds and operational situations. The 4-m deep pond was supplied with pre-screened, untreated wastewater from the nearby treatment plant of Thessaloniki, Greece, initially at a flow rate of 120m3/day and later at a flow rate of 150m3/day. The influent characteristics were BOD5 441 mg O2/L, COD 942 mg O2/L and suspended solids (SS) 574 mg/L. BOD5, COD, and SS concentrations of the pond effluent were reduced by 50%, 53%, and 64%, respectively, in comparison with those of the influent. During the operational period, three distinctly different zones were seen to form within the anaerobic pond: The first zone, which formed at the bottom of the pond, consisted of inert, high-density sludge. The second zone, which formed above this, contained a high concentration of volatile (easily biodegradable) sludge. The third upper zone (supernatant), was a liquid layer low in suspended solids. The accumulation of sludge in the pond followed an annual sinusoidal pattern with high values during winter and low ones during summer due to the increased digestion rate. The maximum high-density sludge height observed was 0.7m, or 2% (14 m3) of the total pond volume. The maximum volatile sludge accumulation reached 3.1 m, or 53% (300 m3) of the pond volume. A seasonal sludge accumulation model, based on the sludge inflow and seasonal digestion rates, was used to simulate the annual fluctuation in accumulation rate for the local (Mediterranean type) climatic conditions. Monthly values of accumulation (or digestion) rate of sludge (K(AS)) were experimentally estimated at specific mean monthly air temperatures and approximated by a regression second degree polynomial equation to be used with the model. The predicted desludging interval for our experimental pond was 3 years.