Abstract: Under the background of climate warming, it is a very realistic and important issue facing China that how to accurately analyze the water loss and gain status of agricultural ecosystems. In order to explore the water loss and gain status at the monthly scale in Songnen Plain, remote sensing data, long-time series meteorological data and soil data were used as base date. In this study, water loss and gain models, trend analysis and wavelet analysis were employed to analyze the spatial and temporal variations of water loss and gain status of Songnen Plain in each month of the crop growth cycle (May to September) from 1961 to 2016. The results show that: (1) From 1961 to 2016, the average potential evapotranspiration during the crop growth cycle was between 432.8 and 605.8 mm in Songnen Plain, and its general variation trend was -0.51 mm·a^-1. The monthly variation trends from May to September were -0.27, -0.16, -0.09, -0.03 and 0.02 mm·a^-1, respectively. (2) The water deficit in the study area was most severe in May and June. The areas with high water deficit were mainly located in the central and western parts of the Songnen Plain, with frequencies between 30% and 50%. The second serious occurred in September, and the areas with high water deficit were located in the northeast, and its frequency was between 20% and 30%. The water deficit was lightest in July and August. (3) According to the periodic regularity and change trend of water loss and gain status, each month showed a clear periodic change on the time scale of 15 years. In addition, from May to July the water loss and gain status showed a periodic downward trend, while for August and September the water loss and gain status showed an upward trend.