This　paper　proposes　a　soft　sensing　method　based　on　a　brain-inspired　modular　stochastic　configuration　network　for　the　real-time　and　accurate　measurement　of　effluent　ammonia　nitrogen　concentration　in　wastewater　treatment　processes.　This　method　first　uses　grey　relational　analysis　to　extract　the　main　variables　affecting　the　effluent　ammonia　nitrogen　concentration.　Then,　based　on　the　fuzzy　clustering　method,　the　task　is　decomposed　to　alleviate　the　complexity　of　each　modeling　task.　After　that,　following　the　principle　of　divide　and　conquer,　a　corresponding　stochastic　configuration　network　sub-model　is　constructed　for　each　sub-task,　and　the　results　of　each　sub-model　are　integrated　to　improve　the　measurement　performance　of　effluent　ammonia　nitrogen　concentration.　Finally,　the　effectiveness　and　accuracy　of　the　proposed　method　are　verified　by　simulation　based　on　actual　water　quality　data.　©　2023　IEEE.