Objective To study the curative effect of dual-target neoadjuvant therapy on human epidermal growth factor receptor-2 (HER-2) positive breast cancer and the changes of magnetic resonance imaging (MRI) parameters.Methods A total of 112 breast cancer patients admitted to the hospital from January 2019 to March 2020 were enrolled and divided into TCH group［trastuzumab (H)， taxotere (T)， carboplatin (C)］ and TCHP group［trastuzumab (H)， pertuzumab (P)， taxotere (T)， carboplatin (C)］ according to different treatment methods， 56 cases in each group. They were followed up for 1 year. MRI and pathological efficacy， and 1-year survival rate between the two groups were compared. MRI examination results in patients with different curative effect were compared. The predictive value of the MRI parameters for pCR of patients with HER-2 positive breast cancer was analyzed by ROC curves. The curves of 1-year survival rate in patients with different curative effect were analyzed by Kaplan-Meier method.Results The response rate of MRI detection in TCHP group was higher than that in TCH group (78.57% vs 60.71%) (P<0.05)， and response rate of pathological detection was higher than that in TCH group (82.14% vs 58.93%) (P<0.05). The differences in apparent diffusion coefficient (ADC)， tumor maximum diameter， EER， TIC and enhancement mode between patients without or with pathological complete response (pCR) were statistically significant (P<005).AUC of ADC combined with tumor maximum diameter and EER for predicting pCR of patients with HER-2 positive breast cancer was 0.891 (95%CI: 0818-0.942).AUC of combined prediction was higher than that of single index (P<0.05). The 1-year survival rate in TCHP group was higher than that in TCH group (92.73% vs 78.78%) ((2=4.192， P=0.041).Conclusion Curative effect of TCHP neoadjuvant therapy is significant on patients with HER-2 positive breast cancer， with high 1-year survival rate and high safety. MRI signs indicate that the combined assessment of ADC， tumor maximum diameter and EER can effectively improve the predictive value for pCR.