This study presents a method for food safety testing based on multi-sensor data fusion, mainly to deal with food safety testing to detect the most common substances, these materials include: formaldehyde, heavy metals lead and cadmium and organ phosphorus and carbamate pesticides. Because it is more sensors, so the detection system will generate a lot of signals. Limited ability of the sensor output signals, these signals are often relatively weak and because the system will be applied to on-site testing and site environment is complex signal is easily influenced by many factors. System design related circuit, using a variety of chips, purify and amplify the signal processing of analog to digital conversion. The final article, but also on data fusion technology relevant circumstances were introduced to the relevant algorithm analysis shows. The experiment data shows that the system has a wide detection range and the configuration is simple with low detection cost which shows the method prompted can meet the engineering requirements of food safety detection.