Seat leakage in valves has become the main concern to the industrial plant. The concerns are not only about the safety of working condition but also the fact that it gives large effect to maintenance costs and production loss. The leakage can either occur quickly or slowly which will make it difficult for personnel to detect the problem on time. This paper discusses the application of acoustic emission method for fault detection in control valve at early stage. During the experimental procedure, the fault was simulated by stages; minor damage, critically damage, and damaged. Data were recorded not only on fully closed position, but also at several openings. Therefore the signal behavior while the valve is operating is known, which contributes towards online fault monitoring. Statistical analysis and Fourier Transform were used to analyze acoustic emission signal from each stage and provide the benchmark values for early fault detection. © 2016 IEEE.