Segmentation of the left ventricular (LV) myocardium in 2D echocardiography is essential for clinical decision making, especially in geometry measurement and index computation. However, segmenting the myocardium is a time-consuming process as well as challenging due to the fuzzy boundary caused by the low image quality. Previous methods based on deep Convolutional Neural Networks (CNN) employ the ground-truth label as class associations on the pixel-level segmentation, or use label information to regulate the shape of predicted outputs, works limit for effective feature enhancement for 2D echocardiography. We propose a training strategy named multi-constrained aggregate learning (referred as MCAL), which leverages anatomical knowledge learned through ground-truth labels to infer segmented parts and discriminate boundary pixels. The new framework encourages the model to focus on the features in accordance with the learned anatomical representations, and the training objectives incorporate a Boundary Distance Transform Weight (BDTW) to enforce a higher weight value on the boundary region, which helps to improve the segmentation accuracy. The proposed method is built as an end-to-end framework with a top-down, bottom-up architecture with skip convolution fusion blocks, and carried out on two datasets (our dataset and the public CAMUS dataset). The comparison study shows that the proposed network outperforms the other segmentation baseline models, indicating that our method is beneficial for boundary pixels discrimination in segmentation.