Price prediction in financial markets has long been a difficult task. However, while many attempts have been made to improve stock market predictability, there are few stud-ies of bond markets. Unlike stocks, most bonds do not trade on exchanges. Consequently, the bond market usually lacks transparency and liquidity, making any estimation of bond prices an especially risky endeavor. Even so, the average daily trading volume of corporate bonds was more than 30 billion dollars. Evidently, the bond market is enormous and the need for improved prediction models that can forecast bond prices and support trading decisions cannot be overestimated. This paper proposes a novel approach to building bond price predictive models based on the technical indicators in financial markets and improv-ing their computing efficiency by applying the machine learning techniques on Apache Spark framework. Our predictive models are constructed in three phases. First, we expand the feature set of each model by transforming the original price time series into a set of technical indicators; the number of features is then reduced by applying dimensionality reduction methods. Second, we employ machine learning algorithms to build predictive models. Finally, we compare the prediction results of different models by evaluating their MAE and RMSE. The data used in this research is a competition dataset from Kaggle containing corporate bond transactions. The experimental results show that our proposed approach considering technical indicators and dimensionality reduction outperforms the baseline for bond price prediction.