<p>Aspect-based sentiment analysis (ABSA) is fine-grained sentiment analysis. Existing ABSA methods usually have the following three drawbacks. (1) Graph-based methods use oversimplified 0/1 adjacency matrices, losing implicit dependency relationship information during training. (2) Syntactic dependency tree-based approaches focus excessively on syntax while disregarding critical semantic features in sentences. (3) Aspect-word interactions often depend on a single pooling method, overlooking the influence of aspect features on the context. These methods often isolate and do not adequately address syntactic or semantic features, insufficiently capturing implicit dependencies and long-range associations within sentences. This paper proposes a hybrid syntactic-semantic enhancement neural network (HSSNN) for aspect-based sentiment analysis to compensate for these deficiencies. This method integrates syntactic and semantic feature extraction through dual-view fusion. Its semantic module leverages a bidirectional long short-term memory network with multi-head self-attention to model global sequential dependencies and extract critical semantic features, enabling fine-grained semantic reasoning. Meanwhile, its syntactic module extracts syntactic information from the dependency parser and captures grammatical relationships between aspect terms and contexts through graph convolutional networks. Experiment results show that the HSSNN model outperforms baseline models on Restaurant, Laptop, and Twitter datasets, averaging an increase of 1.64–1.70% and 1.92–2.83% on the Accuracy and macro-averaged comprehensive scores.</p>