Abstract: In a coal seam gas well, variable speed submersible pumps are used to remove the water from the well bore in order to release the methane gas absorbed in the coal bed and in the water. To optimize gas production, the liquid must be removed at an optimum rate, while still maintaining a certain minimum fluid level within the bore. To achieve this optimum control of the pump, analytical data of liquid flow rate, liquid level trend within the bore, and operating data of the submersible pump is used to control the liquid pump speed and the on-off timing of the liquid pump. High-resolution trending and event-log data provide artificial intelligence control to improve recovery of liquids from a coal seam gas well reservoir, and to reduce the operating cost.