在伺服器製造業面臨品項劇增的環境下，現行庫存管理多以物料金額或數量為導向，難以真實反映大型機構件對倉容造成的「實體空間壓力」，導致臺北港外租倉頻繁空間滿載與生產排程混亂。本研究以Q公司為個案，旨在剖析空間滿載之根本成因並建立具體改善對策。 本研究建構「累積空間占用棧板數」指標量化倉容消耗，並以空間ABC分類法鎖定占用資源之關鍵高占用料號。有別於傳統統計分析，本研究導入「視覺化庫存波形診斷」與ERP系統資料交叉檢核，精準釐清正常循環庫存與異常空間占用（如：關鍵配套料短缺導致主料滯留、批量到貨流速失衡、呆滯舊料未及時處理等）之分野與因果關係鏈。 針對診斷出之異常情境，本研究設計對應的管理觸發點與跨部門防堵機制，將管理防線前推至海運訂艙等不可逆之物流節點。最後，透過長達210天之歷史營運數據進行回溯試算驗證成效。結果證實，及早啟動預警機制並採取對應管理對策，能有效降低異常空間占用，大幅減少總體空間占用棧板數與空間滿載風險，為企業提供一套客觀且可落地的倉容管理解方。;As the server manufacturing industry faces rapid product proliferation， conventional inventory management based on material value or quantity fails to capture the physical space pressure of large mechanical components， leading to frequent overflow at Company Q′s Taipei Port warehouse. This study introduces a "Cumulative Pallet-Days of Space Occupancy" metric and spatial ABC classification to identify key space-consuming materials. A visual inventory waveform diagnosis combined with ERP cross-checking then differentiates normal cycle stock from four abnormal occupancy patterns: critical matching-part shortages， inbound flow-rate mismatches， obsolete material retention， and forecast-gap over-stocking. Based on these scenarios， targeted management triggers and cross-departmental mechanisms are designed to advance the defense line to critical decision points before irreversible space commitment occurs. Effectiveness is verified through historical back testing using 210 days of operational data， showing that timely trigger activation substantially reduces abnormal space occupancy and suppresses peak inventory levels. This study provides an objective， practical warehouse capacity management framework for manufacturing.
