The crowdsourcing-based WLAN indoor localization system has been widely promoted for the ef-fective reduction of the workload from the offline phase data collection while constructing radio maps. Aiming at the problem of the inaccurate location annotation of the crowdsourced samples, the existing invalid access points ( APs) in collected samples, and the uneven sample distribution, as well as the diverse terminal devices, which will result in the construction of the wrong radio map, an effective WLAN indoor radio map construction scheme ( WRMCS ) is proposed based on crowdsourced samples. The WRMCS consists of 4 main modules: outlier detection, key AP selec-tion, fingerprint interpolation, and terminal device calibration. Moreover, an online localization al-gorithm is put forward to estimate the position of the online test fingerprint. The simulation results show that the proposed scheme can achieve higher localization accuracy than the peer schemes, and possesses good effectiveness and robustness at the same time.