In machinery and equipment manufacturing industry, screw is an important part. Not only as a precision screw precision transmission components, but also commonly used for precision measuring devices. Precision screw itself directly affects the accuracy of precision machine tools and precision measuring instruments for transmission, positioning precision and accuracy, and thus the processing and inspection of precision screw is higher demanded. Traditionally, screw inspection is usually done by manual and the quality of inspection is not high, with very low inspection efficiency. To solve these problems, optimized system for screw inspection was designed and implemented. Using the Quartile range method in Statistics to remove outliers, and with the help of kalman filter, systemic error is got rid of. With these methods, raw measurement data is analyzed and processed, and inspection report is generated. The efficiency of screw inspection is highly improved.