In cognitive radio networks, different received signal-to-noise ratios (SNRs) of secondary users (SUs) lead to different degrees of reliability in their local spectrum sensing decisions, whose consideration may greatly affect the sensing capability of a cooperative spectrum sensing system. Cooperative spectrum sensing with traditional hard decisions fusioncan not achive high sensing capabilitydue to allocating equal weights to SUs' decisions. Therefore, a weighted cooperation is considered for cognitive radio networks to enhance the sensing capability. However, in the weighted approch, all SUs are required to report their decisions to the fusion center (FC). This requirement deters the opportunity of efective spectrum usage, especially in relatively fixed environments. The proposed algorithm is a simple advancement towards do resolution. It divides the SUs into "important" and "less important" groups according to their weights. In the first step, SUs of the first group send their local decisions as bit 0(absence) or 1(presence), to FC.According to proposed algorithm, FC sums up the normalizd weights of the same type received bits until one of the summations becomes exceeds 0.5 and declare the presence or absence of a PU. Hence, the local decision of the remaining cooperating users does not affect the final decision. Simulation results show that, compared to weighted spectrum sensing with traditional decisions, proposed weighted spectrum sensing has an acceptable ROC with higher throughput. Keywords: EnergyDetection, Local Decisions, Weighted Cooperative Sensing, Cognitive Radio Networks, Throughput