Nuclear power plant uses the source range channel of the nuclear instrumentation system to monitor the core flux level during refueling and initial startup. Due to the inherent randomness of nuclear reactions, the count rate signal from the source range channel is subject to statistical fluctuations. To mitigate these fluctuations and improve measurement accuracy, this paper analyzes the factors affecting count rate stability and derives a pulse count rate filter model based on optimal estimation theory. The filter’s performance is evaluated through simulations using a random pulse generator to mimic detector output. Results demonstrate the filter’s effectiveness in suppressing statistical fluctuations, thereby enhancing the reliability of neutron flux measurements during critical plant operations.