$\textbf{Objective}$ Develop an automatic diagnostic system which only uses textual admission information from Electronic Health Records (EHRs) and assist clinicians with a timely and statistically proved decision tool. The hope is that the tool can be used to reduce mis-diagnosis. $\textbf{Materials and Methods}$ We use the real-world clinical notes from MIMIC-III, a freely available dataset consisting of clinical data of more than forty thousand patients who stayed in intensive care units of the Beth Israel Deaconess Medical Center between 2001 and 2012. We proposed a Convolutional Neural Network model to learn semantic features from unstructured textual input and automatically predict primary discharge diagnosis. $\textbf{Results}$ The proposed model achieved an overall 96.11% accuracy and 80.48% weighted F1 score values on 10 most frequent disease classes, significantly outperforming four strong baseline models by at least 12.7% in weighted F1 score. $\textbf{Discussion}$ Experimental results imply that the CNN model is suitable for supporting diagnosis decision making in the presence of complex, noisy and unstructured clinical data while at the same time using fewer layers and parameters that other traditional Deep Network models. $\textbf{Conclusion}$ Our model demonstrated capability of representing complex medical meaningful features from unstructured clinical notes and prediction power for commonly misdiagnosed frequent diseases. It can use easily adopted in clinical setting to provide timely and statistically proved decision support. $\textbf{Keywords}$ Convolutional neural network, text classification, discharge diagnosis prediction, admission information from EHRs.