The paper presents the results of experimental studies carried out on the ice of the Gulf of Finland near St. Petersburg in February 2024. The primary objective was to assess the feasibility of using a static station on the ice to correct gravity variations caused by ice cover movements, thereby improving the quality of ground-based gravity survey data obtained from frozen water bodies. The relevance of this work stems from the need for high-precision gravity measurements in ice-covered areas, where standard techniques often yield noisy data. We hypothesized that low-frequency noise from ice oscillations affects both a static and a rover gravity meter simultaneously, or with a minor time shift depending on the distance between them. A field experiment was conducted using two gravity meters (Scintrex CG-5 and CG-6) to analyze the correlation of gravity time series at distances ranging from 0 to 2000 m. Preliminary results confirmed a significant correlation in the low-frequency signal component up to a distance of approximately 500 m (correlation coefficient of 0.73 at 500 m). This validates the proposed correction method in principle but significantly limits its application to detailed, small-scale surveys in shallow water bodies. Filtering and subtracting the low-frequency noise recorded by the static station led to a reduction in the standard deviation of the measurements. Beyond the methodological testing, the study addressed a geological objective: mapping buried valleys that pose a hazard to construction in the city’s coastal reclaimed areas. Observations along one profile revealed a negative gravity anomaly, which is interpreted as a possible continuation of a buried river valley previously identified on the nearby coast. The research is currently limited by the specific conditions of one water body and warrants further investigation under different environmental settings to assess the broader applicability of the method.