Stock forecasting has been a popular and challenging issue in finance and academic researches. Technical and chip indicators are commonly used to forecast the variation of stock index. In addition, the international economic transactions can also influence the variation of stock indices for different countries. Therefore, this research proposes a multiindicator-oriented model based on neural networks. The proposed model combines the technical indicators, chip indicators, international stock market indicators and international exchange rate indicators. The experiments are conducted to find the best parameters on different training set with different lengths of time. The observed multi-indicators provide supportive references for researchers and investors.