A simultaneous methane (CH ₄ ) and carbon monoxide (CO) dual-gas sensor system for coal mine safety applications is highly desirable and yet challenging, from the perspective of complexity and cost. A single distributed feedback (DFB) laser has been utilized in dual-gas sensor to solve the problems in cost and setup complexity. However, there is dual-gas cross-interference to be addressed in single-laser-based dual-gas sensor. To suppress the interference caused by absorption line overlapping, a novel two-step ( ) second harmonic ( ) algorithm based on polynomial fitting ( - -Polyfit) method was proposed for the single DFB laser-based near-infrared dual-gas sensor. A Herriott absorption gas cell was designed with a ~25 m optical path length. Numerical simulations and measurements were carried out to investigate the reported dual-gas cross-interference cancellation method. The signal-to-noise ratio (SNR) with the - -PolyFit algorithm was improved by ~53 dB compared to the scheme without the algorithm. The limit of detection (LoD) of CH ₄ was 0.97 parts per million in volume (ppmv) with a 0.4-s averaging time and it was further decreased to 62 parts per billion in volume (ppbv) with a 224-s averaging time. The CO measurement precision is 0.23 ppmv with a 0.4-s averaging time. With increasing averaging time, a measurement precision of 11 ppbv with a 212-s averaging time was obtained. Field experiments were carried out to evaluate the sensor performance for early fire detection and CH ₄ gas leakage monitoring. The reported novel - -PolyFit algorithm-based dual-gas sensor with no additional requirement on hardware shows enhanced selectivity and antiinterference ability compared to the infrared dual-gas sensor using a traditional sensing architecture with two lasers.