With the rapid increase of the penetration rate of renewable energy sources in the power system, the decrease of system inertia will threaten frequency stability. Moreover, the spatial distribution characteristic of inertia will be more evident and the dispersion of frequency response cannot be ignored. An equivalent inertia estimation method for the system containing wind power considering the influence of the dispersion of frequency response and system partitioning is proposed. First, in order to reduce the influence of the dispersion of frequency response on estimation accuracy, the power system is partitioned based on a spectral clustering algorithm. A frequency similarity index is defined based on Pearson's correlation coefficient to locate the optimal measurement nodes of the area frequency. Then, since the rate of change of frequency (RoCoF) index contains a lot of oscillating components, a numerical integration method based on a swing equation is proposed to estimate the area inertia and system global inertia. Finally, the modified IEEE 10-generator 39-bus system is established in DIgSILENT/PowerFactory to verify the effectiveness of the proposed method. The simulation results show that the proposed method is applicable for the equivalent inertia estimation of a power system with wind power in different scenarios. This work is supported by the National Natural Science Foundation of China (No. 51977128).