The objective of this study was to evaluate a series of mud systems and additives typically used in coalbed methane drilling in terms of formation of an instantaneous filter cake, ability of the coal reservoir to rid itself of the filter cake during production, and overall impact on coal permeability, To achieve this, a series of laboratory tests were conducted initially using artificially cleated gypstone rock (to simulate coal). This was followed by the use of large-diameter coal cores, which, unfortunately, did not allow for the tests to be done under in-situ confining stress conditions. The three mud systems tested against coal (Xantham Gum, HEC and Na-CMC) did not have a negative impact on coal permeability, in contrast to previous laboratory data that showed large decreases. Two fluid loss control additives, which have been used successfully in drilling clastic and carbonate rocks, were also tested using a non-ionic polymer mud system. During simulated drilling, these additives (FLC 2000 (TM) and Q-Stop) were very effective in building a thin filter cake on the coal surface almost instantaneously, to the point that no solids were detected in the downstream fluid accumulator. During simulated production, a small pressure drop was sufficient to remove the filter cake. Coal permeability (to water) returned to its original (pre-test) value, which suggested that there was no permanent permeability damage caused by the two additives. When coal-derived fines were added to the drilling mud in another experiment using the same coal, the near wellbore coal permeability was reduced by 87.5%, indicating severe damage to the cleat system and in agreement with previously reported laboratory data. Following the very good performance of FLC 2000 (TM) and Q-Stop in the laboratory tests, these two additives were then used in field applications. Their presence in the drilling fluid resulted in the successful drilling of 953 m and 1400 m of total horizontal length in the deep Mannville coals in Alberta (at True Vertical Depth of 1400 m and 1150 m, respectively). No borehole instability problems were encountered during drilling of the two horizontal wellbores. The monitored mud losses were low in both cases, with the horizontal well #2 experiencing lower mud loss possibly as a result of the absence of large fractures encountered along the horizontal path. Horizontal well #1 remained stable, which allowed sufficient time to insert a production liner. (C) 2009 Elsevier B.V. All rights reserved.