Coral microatolls are often used to reconstruct the relative sea-level (RSL) along tropical coastlines. They grow at a constant rate, developing each year a growth band that can be observed in their internal stratigraphy. As their development is limited by the water height, they record annual variations of the relative sea-level once they have reached the sea surface. These changes are related to both climate and tectonic, and several criteria are used to decipher both signals. For example, it is commonly accepted that a local signal would rather correspond to a tectonic event, and inversely. However, majority of the criteria such as regrowth of the coral, amplitude of the RSL anomaly or matches with seismic or climatic events catalogs are mainly qualitative and most of the time, incomplete. In our study, we seek to develop a mathematically sound method to separate the climatic signal recorded by a series of microatolls. We focused on the region of the Ryukyus islands in south-west Japan, where the Philippine sea plate plunges under the Eurasia plate. In this area, up to 15 modern and living corals have been collected previously; and their RSL records extend from 1762 to 2018. They extend over 900 km along the subduction zone. Despite the seismic activity of the area, it is possible to infer that any signal common to all microatolls can be considered as climatic. We used a statistical method over the corals dataset to extract a common-mode RSL signal over the island arc. We found a long term sea-level rise for the last 200 years. We further analyze shorter time trends and annual anomalies, and compare our results from the RSL records that include years where only minimum RSL was recorded. Additionally, to refine our method we aim to compare sea level changes recorded by tide gauge in the Ryukyus with the estimates from inferred from the coral microatolls from seismically stable regions in the Pacific Ocean.