Coastal areas host valuable but vulnerable marine ecosystems subjected to increasing anthropogenic pressure and climate change consequences. To assess the impact of these pressures, monitoring programs have proliferated in coastal areas, but most of them follow locally established procedures for quality control (QC). The well-established QC procedure of open ocean data cannot simply be extended to the highly variable coastal area, for which there is the need to develop ad hoc QC approaches. This is particularly crucial for long-term time series, where different instrumentation and analytical methods have been applied over time. This study, based on the biogeochemical dataset collected over 30 years at the LTER MareChiara station (LTER-MC, Gulf of Naples, Mediterranean Sea), addresses potential discrepancies in a long-term dataset, identifying criteria and methods that could also be applied to other coastal datasets. We developed a serial step-wise procedure to characterize the quality of similar to 84,000 data-points, merging statistical tests and expert knowledge. The procedure included nine tests, each addressing potential problems in data generation and management, some of which of general application and others tailored to specific subsets of data. Based on these test, quality flags were assigned to individual data. Critical tests applied to two other independent datasets, showed that the procedure is not dataset dependent. These results contribute to bridge the gap between the need of objective QC criteria and the intrinsic noise of coastal datasets, promoting the discussion on this topic, and improving a proper management and sharing of coastal data.