An overwhelming majority of existing work on soft-decision cooperative spectrum sensing makes the assumption that secondary users (SUs) in cognitive radio networks (CRNs) are stationary. Considering that mobility is an essential property of wireless networks, we analyze the detection performance of soft-decision sensing by mobile SUs. Then, a data fusion rule is proposed aiming at mobile CRNs. Unlike conventional rules, ours needs no a prior information, which makes it more applicable and more suitable to mobile CRNs since prior information like SNR or each SU's location is not available when SUs keep moving. Finally, we analyze the performance of our rule and illustrate its superiority compared to conventional ones. Simulations show that our rule increases detection probability, decreases miss detection probability and error probability on condition that false alarm probability keeps the same, compared with equal gain combination (EGC).