This invention relates to a continuous respiratory rate estimation motion. By a pulse oximetry sensor continue to measure a pulse of heart to produce a photoplethysmography (PPG) and transport this to a processing module. The processing module analyze the PPG to produce one to more respiratory rate. Additionally, a learning module generate learning prediction rate frequency. A decision module analyze the PPG and/or one to more consistency of the respiratory rate. According to the PPG is excellent or bad, or the respiratory rate is consistency or not to adopt different strategy to fusion one to more the respiratory rate and the learning prediction rate frequency, to generate a respiratory rate signal.一種呼吸速率連續評估方法，係由一脈搏血氧感測器持續量測一脈搏，以產生一脈搏血氧飽和度信號並傳輸至一處理模組，再分析脈搏血氧飽和度信號一至多個變化值以產生一至多個呼吸速率，另外由一學習模組產生學習預測速率，由一決策模組分析脈搏血氧飽和度信號品質，或/及一至多個呼吸速率之一致性，根據信號品質好壞，或呼吸速率之一致性與否，依此採用不同策略融合一至多個呼吸速率及學習預測速率，以產生一呼吸速率信號。
