The Tower of London (ToL) is a transformation task extensively used and well-established as a neuropsychological diagnostic tool for assessing human planning ability in clinical and research contexts. Behavioral experiments have recently shown that planning in the ToL is substantially influenced by structural task parameters. This work presents an ACT-R model of the ToL that explains structural influences by using different strategies, whereby, strategy selection depends on visually observable characteristics. Model evaluation was based on a problem selection that accounted for systematic variations of task demands. Based on comparisons with empirically observed planning latencies from previously published data, we argue that task-specific structural characteristics are necessary to explain human planning strategies.