As pollution and climate change intensify, reducing emissions from port-related industries has become a global priority. Although shore-power berths can curb dockside emissions and have attracted substantial investment and policy support, their utilization often remains low, limiting realized environmental benefits. To increase shore-power adoption, this study investigates a dynamic berth pricing and allocation problem, where visiting ships make endogenous berth choices in response to pricing incentives for using shorepower berths. The problem is inherently multi-stage and involves multiple sources of uncertainty, including ships' choice behaviors, arrival times, and berthing durations. To address these challenges, we propose a choice-based P-Queue robust satisficing approach that integrates a multinomial logit model to capture ships' choice behaviors, enabling a tractable reformulation as an exponential cone program. We develop a novel cutting-plane-enhanced binary search algorithm to solve the model and prove its convergence to the optimum. Using operational data from Dalian Port, the proposed approach demonstrates superior out-ofsample performance relative to benchmarks, improving port profit by about 9% and shore-power utilization by 11%. Our solution framework can also be applied to identify the profit-maximizing "sweet-spot" mix of shore-power and conventional berths, providing a implementable decision-support tool for dynamic pricing, infrastructure planning, and policy design.