A spring-composite index (s-CI) is proposed in this study that involves slightly altering the use of the accumulated precipitation from the composite index (CI) comparing the value with other three commonly used indices (standardized precipitation index, SPI; self-calibrated Palmer drought severity index, sc-PDSI; and CI). In addition, the spatial–temporal variation of the s-CI in the Songnen Plain (SNP) was investigated using the Mann–Kendall test and empirical orthogonal function (EOF) methods. The results indicated that the proposed s-CI could identify most drought events in 1990s and 2000s and performed relatively better than SPI, sc-PDSI, and CI in this region. Compared with the other three indices, the s-CI had a higher correlation with relative soil moisture in April and May. The recent spring droughts (2000s) were the most severe in April or May. The weather was drier in May compared with April in the 1980s, whereas the weather was wetter in May than in April in the 1960s and 1970s. Moreover, the spatial patterns of the first EOFs for both April and May indicated an obviously east–west gradient in the SNP, whereas the second EOFs displayed north–south drought patterns. The proposed index is particularly suitable for detecting, monitoring, and exploring spring droughts in the Songnen Plain under global warming.