This paper presents a task scheduling model based on intelligent signal environment cognition is provided to analyze the phased array radar resource scheduling in complex combat environment. The traditional priority of scheduling methods can't be satisfied to the needs in practice. According to the work mode and task of the phased array radar, some complex modulated signal waveform was established and WVD time-frequency transform was used to preprocess the features such as position of geometry center, polarimetric characteristics and et al which are provided to classifier and obtain the task priority. To overcome the weakness of classical scheduling algorithm, a novel task scheduling algorithm is provided which is based on Gaussian Process multitask classification. The method can effectively improve the performance of the radar scheduler and make the task scheduling more feasibly and practical.