Mirzaei, SPrasanna, RTan, M.LHerath, P2026-07-302026-05-212783 – 8862https://rda.sliit.lk/handle/123456789/5120Earthquake Early Warning (EEW) systems built on low-cost MEMS sensors and decentralised processing architectures represent a promising path toward affordable, largescale seismic warning coverage. Yet despite considerable progress at the component level, the field lacks a formal engineering infrastructure capable of binding these components into a coherent, verifiable, and maintainable system. This paper presents work in progress toward filling that gap through the systematic application of Model-Based Systems Engineering (MBSE) to the design of a decentralised EEW framework. Grounded in Design Science Research Methodology (DSRM), the approach uses Systems Modelling Language (SysML) to produce a formal, traceable representation of system architecture and behaviour. We report on three completed stages: a systematic literature review that identified the core research gaps, a component analysis that produced a conceptual framework for the decentralised architecture, and the formalisation of node-level detection and alert logic. The ongoing SysML modelling activity and the path toward simulation-based validation are also described. The central argument is that MBSE is not merely a documentation exercise for EEW research; it is the missing engineering foundation that makes these systems verifiable, scalable, and transferable.enearthquake early warningdecentralised systemsmodel-based systems engineeringMEMS sensorsSysMLdiscrete event simulationConference Paperhttps://doi.org/10.54389/BEOG7633