This design uses optimized hardware circuit, selected device and virtual instrument software platform LabWindows/CVI data processing and analysis functions, digital filtering, nonlinear compensation and multiple acquisition and average processing. The experimental results show that compared with the traditional spring sorter, the relative error of detection is reduced from 1.45% to 0.23%, and the resolution is increased from 0.5N to 0.01N in the range of 0-500N. The overall structure of the detection system transforms the output of the spring sorter measurement sensor through the signal conditioning circuit in the standard input range of the data acquisition card. The A/D of the acquisition card converts the analog quantity into a digital quantity and sends it to the computer, and realizes the acquisition by CVI programming. The data is filtered, multi-point averaged, etc., according to the pre-set qualified range requirements, the D/A corresponding to the acquisition card is started. The control amount of the system is the switch quantity, so the programming control output can be 0V or 5V. The output is triggered by the driving circuit to activate the FET switch to open the corresponding bin. The detection part is the main part of the whole system, and its detection accuracy directly affects the sorting result. The pressure sensor adopts GYL-3 type, the range is 0500N, the accuracy level is 0.05, the power supply is 5V, and its full-scale output is 10mV. The signal amplification adopts three differential precision amplifier circuit, and the integrated operational amplifier chip is determined by repeated comparison and selection to adopt AI4741. The detection system software design collects 50 data from the same spring and sends it to the computer for processing. The digital filtering is realized by CVI programming. The reasonable selection of parameters can effectively eliminate the interference, and the average value of the filtered data is used as the current measurement output. The output of the load cell is a DC signal. After amplification, the detection data is processed by the IVR (Infinite Impulse Response Filter) filter function InvCh-LPF (Chebyshev low-pass filter) provided by CVI. Test results Under the same experimental conditions (same temperature, power supply, environment), the traditional circuit and the designed virtual instrument platform-based detection system are used to measure the 200N force for 10 times. The traditional test data is due to power fluctuations and the noise of the electronic components themselves. And various interference effects, and lack of effective data processing, the measured data fluctuation is 1.45%, the CVI-based detection system uses the powerful functions of the computer to replace part of the hardware, and effectively processes the data, the parameters are adjustable, under the same conditions, measurement The resulting data fluctuated by 0.23%. As a new method to establish measurement and control systems, virtual instrument technology has gradually been applied to many fields such as aerospace, petroleum, and fault diagnosis. Based on the improvement of the hardware circuit of the spring sorter, the paper establishes the measurement and control system and performs data processing based on the LabWindows/CVI software platform. The experimental and field operation results show that the accuracy and reliability of the spring sorter can be effectively improved. The relative fluctuations in the output of the sorter currently in use in China are reduced from 1.45% to 0.23%. After being used by a domestic textile machinery factory, it is believed that the fault-free running time of the spring automatic sorting instrument has reached three years, which is more than 6 times longer than domestic similar products.