During prolonged extreme meteorological events such as cyclones, low-level rapids, thunderstorms, etc., the output of wind farms may undergo noticeable fluctuations within short timeframes, leading to wind power ramp events (WPREs). This paper proposes an optimization method for allocating electrochemical energy storage (EES) to tackle WPREs considering the correlation among offshore wind farms. Firstly, the correlation of offshore wind farms’ output is verified and analysed. Multiple scenarios are sampled using Latin Hypercube Sampling (LHS) and then rearranged using Cholesky decomposition ordering method (CDOM) to obtain output scenarios that incorporate the varying correlations among adjacent wind farms. Secondly, an improved multi-parameter segmentation algorithm is proposed to detect WPREs for the different correlation scenarios and assess the demand of EES for WPREs. Following this, a bi-level optimization model of EES coordinated with thermal power units (TPUs) for WPREs is formulated. Finally, through an example analysis, the effectiveness of the proposed model is verified. The simulation results show that stronger correlation among wind farms lead to fewer WPREs and lower requirements for TPUs and EES capacity, thereby reducing the investment and operation costs.