Paper chip is an efficient and inexpensive device for pesticide residues detection. However, the reasons of detection error are not clear, which is the main problem to hinder the development of pesticide residues detection. This paper focuses on error analysis for pesticide detection performed on paper-based microfluidic chip devices, which test every possible factor to build the mathematical models for detection error. In the result, double-channel structure is selected as the optimal chip structure to reduce detection error effectively. The wavelength of 599.753 nm is chosen since it is the most sensitive detection wavelength to the variation of pesticide concentration. At last, the mathematical models of detection error for detection temperature and prepared time are concluded. This research lays a theory foundation on accurate pesticide residues detection based on paper-based microfluidic chip devices.