Jet identification is one of the fields in high energy physics that deep learning has begun to make some impact. More often than not, convolutional neural networks are used to classify jet images with the benefit that essentially no physics input is required. Inspired by a recent paper by Datta and Larkoski, we study the separation of quark/gluon-initiated jets based on fully-connected neural networks (FNNs), where expert-designed physical variables are taken as input. FNNs are applied in two ways: trained separately on various narrow jet transverse momentum $p_{TJ}$ bins; trained on a wide region of $p_{TJ} \in [200,~1000]$ GeV. We find their performance are almost the same, and the larger $p_{TJ}$ the better. Comparing with results from deep convolutional neural networks, our results are comparable for low $p_{TJ}$, and even slightly better than those for high $p_{TJ}$. We also test the performance of FNNs with full set or different subsets of jet observables as input features. The FNN with one subset, consisting of fourteen observables, shows nearly no degradation of performance. This indicates that these fourteen expert-designed observables may have captured most of the information useful for the separation of quark/gluon jets.