A novel touchscreen-based lock pattern recognition method has been addressed in this paper. In most cases the lock-patterns are drawn on a grid of particular size, and commonly it is of size \(3 \times 3\) . For grid based patterns, it is easy to measure similarity by following the sequence of the grid vertex labels but the patterns are easily imitable. In this paper, we propose a method for grid-free lock pattern recognition utilizing Prufer sequences. We introduce a pattern graph vertices labeling method that effectively encodes the spatial relationships among the dominant points lying on the pattern. By generating Prufer sequences from labeled pattern graph vertices, we create a compact yet informative representation of the patterns. The distance between two code sequences serves as a key metric for measuring the matching score between input and reference pattern. The Levenshtein distance has been used by us for matching score computation which not only considers the spatial arrangement of the dominant points but also takes into account their sequential order, enhancing the accuracy of recognition. We have gathered data from touchscreen interactions and curated a custom dataset. Our experimental results demonstrate strong performance across various patterns. Here, an input shape from the user is drawn on the touchscreen and subsequently the similarity of this input is measured with already set pattern for the user. As, we are intending for grid-free patterns, similar or almost similar inputs (shape-wise) may be accepted as a valid input for the user. This paper shows how efficiently the proposed method can validate grid-free patterns. The grid-free patterns are easy to remember and draw on the touchscreen; and also the proposed methodology can avoid the threat of being imitated.