Because it has many advantages such as rapidity and accuracy, nucleic acid detection is applied to infectious disease diagnosis more and more. An automatic integrated nucleic acid detection system based on real-time PCR is developed by our research group to conduct point-of-care testing of infectious pathogens. The home-made detection system collects fluorescence data in each PCR cycle through an integrated dual-channel fluorescence detection module and then real-time fluorescence curves are drawn by the software, which can tell the results of the diagnostics after some processing and analysis. However, owing to the disturbance of the environment or the imperfect of nucleic acid extraction before PCR, the fluorescence curves sometimes may contain several abnormal points. For the purpose of enhancing its ability to deal with these iffy curves and improve the accuracy of the testing results, in this study, the SDM-based qPCR data processing algorithm was studied and 11 groups of qPCR data that have different flaws from the clinical samples detected by this system were chosen to prove the practicability of the method. In comparison with the conventional threshold-based method, the Cq values calculated by the SDM-based method were more close to the actual values, meaning it can overcome the shortcomings of the conventional methods such as being unable to accommodate noise and being unable to avoiding abnormal data. With the improvement of this data processing algorithm, the stability of our system and the reliability and accuracy of the results are greatly improved.