Abstract: Based on station observations and high resolution gridded precipitation data, the relationship between spring sensible heat in the Tibetan Plateau (STPSH) and summer precipitation in eastern China (SPEC) is investigated in terms of decadal change and interannual variability by using the maximum covariance analysis. Further attempt has been made to establish a statistical model for forecasting the SPEC using the timescale decomposed regression approach. Results indicate that for the decadal component, a significant correlation exists between the STPSH and SPEC in most part of eastern China in June, July and August with the explained variance fractions of 75.6%, 99.9% and 79.7%, respectively. For the interannual component, however, the significantly correlated regions are distributed in southern China, the coastal area of northern China, and the Yangtze-Huai River valley in June; in July, the correlated areas are located over the southwestern part of South China, the Yangtze River valley, the southeastern part of Northeast China and the middle-lower reaches of Yellow River; high correlation is found over Northeast China and the western part of South China in August. The explained variance fractions are 42.7%, 43.4% and 32.0%, respectively. The explained variance analysis and the hindcast examination suggest that the best prediction skill of this model occurs in most part of eastern China in July. The areas with high predictability are the southern region of the Yangtze River in June, and northeastern China and western part of South China in August. The model can reasonably describe the relation between the STPSH and SPEC and quantitatively forecast local precipitation in June, July and August. Therefore this model might be used for short-term operational climate prediction.