In this work, we propose a restricted access points (AP) based dynamic channel estimation for a sixth-generation (6G) cell-free massive multiple input multiple output (CF-mMIMO) system. In CF-mMIMO network architecture, a large number of APs are distributed over a wide area and connected to a centralized processing unit (CPU) through a lossless infinite capacity to serve a smaller number of user devices (UDs) simultaneously. The channel state information (CSI) between is active user and APs is an important parameter that needs to be estimated for establishing an efficient communication link among them. Thus, in this work, we propose a restricted APs based channel estimation for the channel between UD to APs at CPU in CF-mMIMO system. A correlated Rayleigh fading channel with spatial correlation is considered between UDs and APs. A minimum mean square error (MMSE) based channel estimation with pilot contamination is carried out for the proposed CF-mMIMO system. An analytical closed form expression is derived for estimated channel for the proposed restricted APs based 6G CF-mMIMO system. A MATLAB based thorough simulation analysis is carried out and the obtained result is compared with the state-of-art small cell cellular technique. It has been shown that the proposed channel estimation technique gives the near optimal performance in CF-mMIMO system.