Harvesting renewable energy from environment sources can resolve the battery issue for systems composed of mobile terminals and enable long-time autonomous deployment. However, simply applying renewable energy to these systems will usually result in low system performance, not only because the available energy harvested from renewable sources is non-deterministic but the energy level is also low. Furthermore, mobile devices could be power-consuming and their power budget may exceed the available power level. In this paper, we propose an adaptive tracking channel control technique for renewable energy powered GNSS receivers deployed for autonomous field applications. First, we formulate the tracking channel control problem and introduce an improved greedy-based technique that dynamically adjusts signal tracking channels according to the signal propagation channel status and available energy level to attain better positioning time coverage. Then, we introduce a model-free reinforcement learning based technique to further improve the GNSS time coverage by optimizing the control policy. Finally, the evaluation results show that the proposed techniques can achieve significant improvement in positioning time coverage over the existing techniques.