Abstract: To reduce the speed fluctuations in the reflow soldering process， an optimization scheme based on the improved automatic process control (APC) was proposed. The output characteristics were monitored during the production process. When the characteristic value exceeded the predetermined value， APC policy was started to adjust the initial setting parameters and BP neural networkgenetic algorithm technology was used to achieve the optimization process. The Taguchi mass loss function of the product was defined and the Taguchi process capability index was calculated for evaluation. The test results showed that， compared with traditional APC scheme， the improved scheme ensures the process stability， decreases the adjust frequency and improves the optimization speed.

Key words: reflow soldering, automatic process control, parameter optimization, mass loss function, process capacity index