The accounting of Gross Ecosystem Product (GEP) has been conducted in many cities across the country, which has been included in the performance assessment indicators. But the relevant ecosystem services in GEP accounting are mostly based on actual precipitation data and multi-year average precipitation data, which makes it difficult to reflect the actual changes in ecosystem quality and quantity when analyzing inter-annual GEP changes, thus affecting the application of GEP management. Taking Shenzhen as a case, this paper proposes the comparable precipitation conditions that are more representative of the generalized characteristics of local precipitation, and a method of determination based on principal component analysis. This paper calculates the GEP for two adjacent years under comparable and actual precipitation conditions, for a total of three scenarios, and analyses the composition of GEP under different scenarios and the ability of precipitation factors to influence inter-annual GEP variability. The study found that:(1) 2019 was a comparable precipitation year in Shenzhen and was consistent with the precipitation situation described in the Shenzhen Climate Bulletin between 2008 and 2020; (2) The four ecosystem service values associated with precipitation (Sedimentation reduction, Non-point resource pollution retention, Water conservation and Flood reduction) showed an inter-annual increase under comparable precipitation conditions, while the values of Sedimentation reduction and Non-point resource pollution retention increased and the values of Water conservation and Flood reduction decreased under actual precipitation conditions; (3) Precipitation factors significantly affected the interannual variation of GEP, and the variation of GEP change in Shenzhen from 2019 to 2020 under different precipitation conditions was as high as RMB 2.478 billion, while comparable precipitation conditions could effectively reduce the influence of non-anthropogenic factors on GEP change; (4) In the process of calculating multi-year average precipitation, the "kurtosis" of the data would be artificially reduced, resulting in the results not representing the typical climate characteristics of the region; (5) Under comparable precipitation conditions, the inter-annual GEP variation characteristics better reflected the actual variation of ecosystem quality and quantity. The results of the study may provide new ideas and methods for future GEP accounting for management applications.