Fuzzy clustering has received a lot of attention in recent years because of its ability to handle overlapping data sets in real-world applications. This paper presents a new fuzzy clustering technique by combining border fuzzy c-means (B-FCM) and semi-supervised collaborative fuzzy clustering ( \(S^2\) CFC) for classifying land cover from remote sensing imagery. It is called the semi-supervised border fuzzy clustering approach based on the collaborative border fuzzy technique (B- \(S^2\) CFC). The proposed method adopts a significantly different approach by embedding border information and collaborative information into the clustering process to handle data samples at the borders of clusters, which is relevant for any remote sensing data imbued with noise and outliers. This means that it efficiently solves the problem of border samples of the clusters, which restrains time intake through effective initial placement of the centers of the clusters based on an understanding of the distribution of the data. Land cover classification experiments from remote sensing image data show that B- \(S^2\) CFC is significantly better than the other methods regarding the results and time used to compute. In addition, the proposed method has proven to be helpful in combating the effects of vague cluster boundaries and noise, which are among the common problems in remote sensing image classification.