Abstract-Gastroesophageal reflux disease (GERD) is one of the most prevalent gastrointestinal diseases. It is characterized by excessive reflux of gastric content (acid, pepsin, etc.) into the esophagus causing symptoms (heartburn, acid regurgitation, etc.) and mucosal inflammation and injuries. GERD occurs when the lower esophageal sphincter (LES) has a low resting pressure and stomach contents leak back, or reflux, into the esophagus. Therefore, the accurate measurement of the LES pressure is of great importance for the diagnosis of GERD. The LES pressure signal, involving severe respiratory contamination and motion artifacts, demands specific capabilities not provided by conventional data analysis methods. Recently, local regression has proved to be a very attractive technique to the nonparametric regression in statistics. In this contribution we apply the ideas of local regression to develop strategies for selecting smoothing parameters of local linear squares estimators, and present its application on the extraction of the LES pressure in GERD. The results from both extensive simulations and real data demonstrate the ability of local regression to characterize the LES pressure, which is consistent with the clinical observation.