The purpose of this study was to investigate the therapeutic efficacy and prognosis of serum HER2 (sHER2) in patients with advanced breast cancer. We analyzed the sHER2 levels of 200 patients with advanced breast cancer receiving first or second line treatment, the tissue HER2 (tHER2) level was also analyzed. Indicators of therapeutic efficacy and prognosis were objective response rate (ORR), disease control rate (DCR), and time to progression (TTP). The baseline sHER2 level was high in 132 patients and low in 68 patients. The high level of sHER2 is correlated with molecular subtype (p=0.016), visceral metastasis (p<0.001), liver metastasis (p<0.001), tissue HER-2 (tHER2) (p=0.001), and, among tHER2-low tumors (59 patients), the baseline sHER2 high level was associated with a higher proportion of brain metastasis. The ORR of patients with baseline sHER2 high level is higher than those with baseline sHER2 low level (p=0.026). The TTP of patients with baseline sHER2 low level is longer than the patients with baseline sHER2 high level (p=0.024). For patients with baseline sHER2 high level, a significant decrease in sHER2 after two cycles of treatment indicates higher ORR, DCR, and an extension of TTP. After multiple cycles of treatment, for patients with tHER-2 positive and baseline sHER2 high level, the DCR in the sHER2 decrease in the negative group was higher than that in the continuous positive group (p=0.037). Patients with a rapid decline type of sHER2 dynamic change curve had higher ORR and prolonged TTP compared with patients with other types of sHER2 dynamic change curve. There is no correlation between OS and sHER2 levels. Our study showed that patients with advanced breast cancer had a high level of sHER2 at recurrence, regardless of whether they are tHER2 positive or negative. Dynamic detection of sHER2 can help predict therapeutic efficacy and prognosis, regardless of whether tHER-2 is positive or negative.