The probability of detection, probability of false alarm, and signal-to-noise ratio (SNR) are three core indicators of the detection system for infrared point targets. Under the criterion of constant false alarm detection, this study analyzed the relationship between probability of detection, probability of false alarm, and signal-to-noise ratio(SNR) using single frame detection method, also called single threshold detection for infrared point target systems. This study focused on the analysis and calculation of the selection strategy of the second threshold under a multi-frame accumulation detection method (called the double threshold detection algorithm) and achieved a higher probability of detection and lower probability of false alarm with a lower SNR. A good solution was also found for the problem of detecting high-value targets with high threat and cost for false and missed alarms.