Band-Stop filter is one of the key components in smart grids, wireless communications systems, radar, and signal processing. Due to the ability of suppressing signals at specific frequencies, these filters enable the rejection of unwanted signals in a particular frequency band in a crowded microwave spectra. Specifically, such filters can be utilized to mitigate the effect of harmonics in smart grids. The traditional band-stop filters suffer from lack of high suppression, reconfigurable bandwidth and frequency agility. This paper presents a novel approach to realize reconfigurable and flexible band-stop filters using the computing-in-memory (CIM) technique. The proposed scheme consists of a low-pass filter and a high-pass filter, with their corresponding coefficients programmed into an array of memristive devices. As a result, reconfigurable and low-complexity band-stop filters can be implemented, which support the realization of arbitrary filter impulse responses and high-order filters.