<p>Cluster-weighted models (CWMs) are a fundamental technique in model-based clustering, widely used for capturing complex relationships in regression data. We introduce a novel class of CWMs to handle functional data by projecting functional predictors onto a wavelet basis. This projection simplifies the new regression model into a structure similar to classical CWMs. However, given that the number of predictors often exceeds the number of observations, we assume a latent factor structure for the projected predictors within each mixture component. By imposing constraints on the covariance matrices of the projected data, we develop a family of parsimonious models. This approach leads to the functional cluster-weighted factor analyzers (FCWFA) model. Model parameters are estimated using a maximum likelihood approach via an alternating expectation-conditional maximization (AECM) algorithm. The performance of FCWFA is evaluated through simulations and real-world datasets, demonstrating its superiority over established mixture-based methods.</p>