<p>As a key component of magnetorheological polishing equipment, the reliability of peristaltic pump hose directly restricts the effective operation of the whole equipment. Unfortunately, the hose has many failure modes in actual engineering, including fatigue, wear, tear and burst blasting. Meanwhile, these failure modes often have high correlation, which brings great challenges to the reliability analysis of magnetorheological polishing equipment. For that reason, this paper proposes an active learning reliability analysis method combining multiple response Gaussian process (MRGP) and Monte Carlo simulation (MCS) to deal with the correlation between failure modes. Herein, MRGP model is used to describe the correlation between failure modes of peristaltic pump hose, and a surrogate model of limit state functions of hose is correspondingly constructed. Combined with active learning strategy, the MRGP surrogate model of hose is updated iteratively until the convergence is satisfied. On this basis, the reliability of the peristaltic pump hose is analyzed by MCS. Finally, a numerical example is used to verify the effectiveness of the proposed method which is further applied into the reliability analysis of the peristaltic pump hose.</p>