Plastic film cultivation technology, which was hailed as “the second revolution of agricultural production”, had the effects of increasing temperature, conserving soil moisture, suppressing weed and boosting crop yield. However, the major ingredients of agricultural plastic mulching film are polyolefin compounds which are degradationresistant. The mechanization of residual film recycling is an effective approach for solving pollution of residual films. The major operation procedures of residual film recycling include releasing film, collecting film and removing film, and unloading and releasing film operations are the crucial link of recycling residual films. Concerning existing releasing equipment of residual film collector had the problem of low reliability and film releasing rate, a runnertype film releasing equipment was designed. Through analysis on the film releasing mechanism of this equipment, its structural characteristics and dimension parameter were determined. Taking the space between film releasing shovels, forward speed and penetrating angle as influence factors and film releasing rate as response value, the influence of each factors on film releasing rate was calculated based on the regression analysis and response surface analysis with Design-Expert 8.0.2 software. The contribution of each factor to the influence on film releasing rate from big to small were penetrating angle, space between film releasing shovels and forward speed. Three-factor quadratic regression models between film releasing rate and penetrating angle of film releasing shovels, forward speed and space between film releasing shovels were established. Optimizing calculation of structure and working parameters of film releasing shovels was conducted through calculation method of nonlinear optimization. The results showed that when the space between film releasing shovels was 220mm, forward speed was 1.0m/s and penetrating angle was 30°, the theoretical maximal film releasing was 93.8%. Proof test indicated that the film releasing rate with these parameters was 91.3% and the error between theoretical value and trial value was 2.5%, which verified the correctness of regression model.