Rapid,　accurate　and　reliable　measurements　of　biological　oxygen　demand　(BOD)　are　a　key　basis　for　monitoring　and　controlling　wastewater　treatment　processes　(WWTP).　A　kind　of　soft　measurement　based　on　the　dynamic　neural　network　(DNN)　is　proposed　in　this　paper,　which　can　be　used　to　monitor　and　model　the　important　parameters　of　the　wastewater　treatment　process　on-line.　The　main　parts　of　the　soft　measurement　are:　soft　measurement　model　design　and　dynamic　neural　network　design.　The　results　of　the　simulations　demonstrate　that　this　soft　measurement　can　measure　the　BOD　concentration　on-line,　it　is　provided　with　real-time　ability,　good　stability,　and　high　precision　and　so　on.　©2009　ACA.