Objective To establish and evaluate the application of direct mercury analysis for the determination of total mercury in food. Methods A direct mercury analysis method was developed based on catalytic pyrolysis-gold amalgamation atomic absorption spectroscopy. By stepwise optimization of pyrolysis temperatures (150 °C/550 °C for low-concentration ranges and 200 °C/650 °C for high-concentration ranges) and carrier gas flow rates (150-200 mL/min), plant-derived samples (e.g., rice, jujube), animal-derived samples (e.g., snakehead, fish meal), and 14 kinds of food samples were analyzed. The method’s detection limit, accuracy, stability and spike recovery were evaluated to assess its performance and matrix applicability. Results The method established dual linear calibration curves of 0-10 ng and 25-200 ng, with correlation coefficients of 0.9993 and 0.9998, respectively. The limit of detection and limit of quantitation were 0.0002 mg/kg and 0.0005 mg/kg, respectively. Spike recoveries ranged from 90% to 116%, the intra-day and inter-day repeatability relative standard deviation was less than or equal to 6.5% and less than or equal to 7.7%, respectively. The method was systematically validated for complex matrices, including plant-derived, animal-derived, and processed foods, with all results complying with the GB 2762-2022 National food safety standard-Limits of contaminants in foods. Conclusion The proposed method eliminates the need for chemical digestion in sample pretreatment, significantly shortens the analytical time, and prevents mercury volatilization losses. It overcomes the limitations of single-matrix analysis and achieves rapid total mercury detection in diverse complex food matrices, thereby providing an efficient and reliable technical solution for food safety risk monitoring and regulatory compliance.